In the remote sensing domain, it is crucial to complete semantic segmentation on the raster images, e.g., river, building, forest, etc, on raster images. A deep convolutional encoder--decoder (DCED) network is the state-of-the-art semantic segmentation method for remotely sensed images. However, the accuracy is still limited, since the network is not designed for remotely sensed images and the training data in this domain is deficient. In this paper, we aim to propose a novel CNN for semantic segmentation particularly for remote sensing corpora with three main contributions. First, we propose applying a recent CNN called a global convolutional network (GCN), since it can capture different resolutions by extracting multi-scale features from different stages of the network. Additionally, we further enhance the network by improving its backbone using larger numbers of layers, which is suitable for medium resolution remotely sensed images. Second, "channel attention'' is presented in our network in order to select the most discriminative filters (features). Third, "domain-specific transfer learning'' is introduced to alleviate the scarcity issue by utilizing other remotely sensed corpora with different resolutions as pre-trained data. The experiment was then conducted on two given datasets: (i) medium resolution data collected from Landsat-8 satellite and (ii) very high resolution data called the ISPRS Vaihingen Challenge Dataset. The results show that our networks outperformed DCED in terms of $F1$ for 17.48% and 2.49% on medium and very high resolution corpora, respectively.

Pedestrian object detection plays a significant role in intelligent systems such as intelligent traffic and monitoring. Traditional machine learning methods on pedestrian detection have shown various drawbacks, \eg{} low accuracy, slow speed, \etc The Convolutional Neural Network (CNN) based object detection algorithms have demonstrated remarkable advantages in the field of pedestrian detection. However, the mainstream CNNs still face the problems of slow speed and low detection accuracy, especially on small and occluded targets from aerial perspective. In this paper, we propose Multi-Scale Attention YOLO (MSA-YOLO) detection algorithm to address the above issues. MSA-YOLO includes a Squeeze, Excitation and Cross Stage Partial (SECSP) channel attention module for CNNs to extract richer pedestrian features with a small number of extra parameters. It also contains a multi-scale prediction module to capture the information among different pedestrian scales, which can recognize the small objects with higher accuracy and significantly reduce the missed detection. To sufficiently evaluate our proposed model, we manually collect and annotate a new benchmark dataset, Luoyang Pedestrian Dataset \footnote{The dataset can be downloaded through this anonymous link: \url{https://drive.google.com/drive/folders/13po1FX7Qk5RDgb60-dzOi74cq_kqDJtM?usp=sharing} }, which has much more sample annotations, features, scenes and image view angles than the existing benchmark datasets. In addition, the images in our dataset have higher resolution than most of benchmark pedestrian detection datasets, which can provide more detailed features of pedestrians and thus improve the model performance. When tested on Luoyang Pedestrian Dataset, our proposed MSA-YOLO algorithm significantly outperform the most commonly used baseline models with almost the same model size. This shows the efficiency of our proposed model. (The code and new dataset will be released to the public later.)

Ion channels are integral membrane proteins that facilitate the movement of ions across cell membranes, playing a key role in a range of biological processes. The high cost and time required for wet lab experiments to characterize ion channels has spurred the development of computational methods for this purpose. In our previous work, we demonstrated the effectiveness of protein language models for ion channel prediction, using a logistic regression classifier to distinguish ion channels from non-ion channels (TooT-BERT-C) and transporters from non-transporters (TooT-BERT-T). In this study, we build upon this approach by using a combination of classical machine learning classifiers and a Convolutional Neural Network (CNN) with fine-tuned representations from ProtBERT, ProtBERT-BFD, and MembraneBERT to discriminate ion channels from non-ion channels. The results of our experiments demonstrate that TooT-BERT-CNN-C, a combination of the representations from ProtBERT-BFD and a CNN, outperforms existing state-of-the-art methods for predicting ion channels, with a Matthews Correlation Coefficient (MCC) of 0.86 and an accuracy of 98.35% on an independent test set.

Partial differential equations (PDEs) are used to describe a wide range of phenomena, such as heat transfer, fluid dynamics, and quantum mechanics. By solving PDEs, we can ob- tain insights into the behavior of the system and make predictions about its future evolution. Conventional numerical methods for obtaining the approximate solutions of PDEs may re- quire extensive computational resources and time, especially for complex PDEs and large- scale problems. The recently developed physics-informed neural network (PINN) has shown promise in a variety of scientific and engineering fields by incorporating physical laws into the loss functions of the neural network (NN). In addition, the training of PINN does not require ground truth data, but it has poor generalization ability to unseen domains. On the other hand, a convolutional neural network (CNN) has fast inference and better generalization ability, but it requires a large amount of training data. Taking the advantages of PINN and CNN by using Legendre multiwavelets (LMWs) as basis functions, we introduce a new method to approach the PDEs in this paper, namely Physics-Informed Legendre Multiwavelets CNN (PiLMWs-CNN), in order to continuously approximate a grid-based state representation that can be handled by a CNN. PiLMWs-CNN enable us to train our models using only physics-informed loss functions without any pre- computed training data, simultaneously providing fast and continuous solutions that gener- alize to previously unknown domains. In particular, the LMWs can simultaneously possess compact support, orthogonality, symmetry, high smoothness, and high approximation order. Compared to orthonormal polynomial (OP) bases, the approximation accuracy can be greatly increased and computation costs can be significantly reduced by using LMWs. We applied PiLMWs-CNN to approximate the damped wave equation, incompressible Navier-Stokes (N- S) equation, and two-dimensional heat conduction equation. The experimental results show that this method provides more accurate, efficient, and fast convergence with better stability when approximating the solution of PDEs.

Post-training quantization is pivotal in deploying convolutional neural networks for mobile applications. Block-wise reconstruction with adaptive rounding, as employed in prior works like BrecQ and Qdrop, facilitates acceptable 4-bit quantization accuracy. However, adaptive rounding is time-intensive, and its constraint on weight optimization space curtails the potential for quantization performance. The optimality of block-wise reconstruction hinges on the quantization status of subsequent network blocks. In this investigation, we delve into the theoretical underpinnings of the limitations inherent in adaptive rounding and block-wise reconstruction. Our exploration leads to the development of a post-training quantization methodology, designated as AE-Qdrop. This algorithm operates in two distinct phases: block-wise reconstruction and global fine-tuning. The block-wise reconstruction phase introduces a progressive optimization strategy, superseding adaptive rounding, which not only augments quantization precision but also significantly improves quantization efficiency. To mitigate the risk of overfitting, we introduce a random weighted quantized activation mechanism. During the global fine-tuning phase, we account for interdependencies among quantized network blocks. The weight of each network block will be corrected with logit matching and feature matching. Extensive experiments validate that AE-Qdrop achieves high-precision and efficient quantization. For instance, in the case of 2-bit MobileNetV2, AE-Qdrop outperforms Qdrop by achieving a 6.26% enhancement in quantization accuracy and quintupling the quantization efficiency.

Landslide is one of the most common geological disasters in China, which is characterized by suddenness and uncertainty, and it is difficult to realize accurate identification, early warning and forecasting of landslide disaster by conventional means. With the development of high-resolution remote sensing satellites and InSAR surface deformation monitoring technology, the traditional means of landslide monitoring data sources are limited, and there is a lack of effective methods to excavate the characteristics of the spatial distribution of landslide hazards and their triggering factors and other issues. In this paper, the area extending 10 km outside the VII isobar of the Gengma earthquake is taken as the study area, and 13 evaluation factors are screened out by integrating the factors of InSAR surface deformation, topography and geological environment, and the Bayesian Optimized Convolutional Neural Network (BO-CNN) is used for the evaluation of landslide susceptibility, and the BO-RF and PSO-SVM models are selected for the comparative analysis. The model accuracy evaluation was carried out by three indexes: ROC curve, AUC value and FR value, in which the ROC curves of PSO-SVM, BO-RF and BO-CNN were all close to the upper-left corner of the corner, and the AUC values were 0.9388, 0.9529, and 0.9535, respectively, and the FR value of landslide in the high susceptibility area of BO-CNN was as high as 14.9, and was higher than that of PSO-SVM and BO-RF, respectively. SVM and BO-RF model is 4.55 and 3.69 higher, the experimental results show that the BO-CNN model used in this paper has a better effect in landslide susceptibility evaluation, and the research results of the local government's disaster prevention and mitigation measures are of great significance.

The correct classification of defects originating from partial discharges (PD) in medium-voltage (MV) switchgears with air insulation (AIS) remains a challenging research topic for scientists worldwide. In this article, the authors simulated four possible defects occurring in the power industry, including one that is a simultaneous combination of two commonly ones. In addition, the correctness of the algorithm was checked by adding a classification class without any fault. The measurement signals were recorded with TEV sensors. The effectiveness of various hy-brid-connected neural networks was tested and discussed: GoogleNet and SqueezeNet based on spectrograms, SAE with FNN, 2D-CNN with LSTM, and hybrid AE combined with CNN and LSTM. The highest effectiveness – approximately 97% – was demonstrated by the GoogleNet and SqueezeNet networks. The research results are expected to form the basis for the development of a universal and wireless capacitive sensor for monitoring the level of PD in switchgears.

Recently, the number of vehicles on the road, especially in urban centers, has increased dramatically due to the increasing trend of individuals towards urbanization. As a result, manual detection and recognition of vehicles (i.e., license plates and vehicle manufacturer) become an arduous task and beyond human capabilities. In this paper, we have developed a system using transfer learning-based DL techniques for automatic identification of Jordanian vehicles. The YOLOv3 (You Only Look Once) model was re-trained using transfer learning to accomplish the license plate detection, character recognition, and vehicle logo detection. While VGG16 (Visual Geometry Group) model was retrained to accomplish the vehicle logo recognition. To train and test these models, four datasets have been collected. The first dataset consists of 7,035 Jordanian vehicle images, the second dataset consist of 7,176 Jordanian license plates, and the third dataset consists of 8,271 Jordanian vehicle images. These datasets have been used to train and test the YOLOv3 model for Jordanian license plate detection, character recognition, and vehicle logo detection, respectively. While the fourth dataset consists of 158,230 vehicle logo images used to train and test the VGG16 model for the vehicle logo recognition. Text measures were used to evaluate the performance of our developed system. Moreover, mean average precision (mAP) measure was used to evaluate the YOLOv3 model of the detection tasks (i.e., license plate detection and vehicle logo detection). For license plate detection, the precision, recall, F-measure, and mAP were 99.6%, 100%, 99.8%, and 99.9%, respectively. While for character recognition, the precision, recall, and F-measure were 100%, 99.9%, and 99.95%, respectively. The performance of license plate recognition stage was evaluated by evaluating these two sub-stages as a sequence, where the precision, recall, and F-measure were 99.8%, 99.8%, and 99.8%, respectively. Furthermore, for vehicle logo detection, the precision, recall, F-measure, and mAP were 99%, 99.6%, 99.3%, and 99.1%, respectively, while for vehicle logo recognition, the precision, recall, F-measure were 98%, 98%, and 98%, respectively. The performance of vehicle logo recognition stage was evaluated by evaluating these two sub-stages as a sequence, where the precision, recall, and F-measure were 95.3%, 99.5%, and 97.4%, respectively.

Object segmentation on remotely-sensed images: aerial (or very high resolution, VHS) images and satellite (or high resolution, HR) images, has been applied to many application domains, especially road extraction in which the segmented objects are served as a mandatory layer in geospatial databases. Several attempts in applying deep convolutional neural network (DCNN) to extract roads from remote sensing images have been made; however, the accuracy is still limited. In this paper, we present an enhanced DCNN framework specifically tailored for road extraction on remote sensing images by applying landscape metrics (LMs) and conditional random fields (CRFs). To improve DCNN, a modern activation function, called exponential linear unit (ELU), is employed in our network resulting in a higher number of and yet more accurate extracted roads. To further reduce falsely classified road objects, a solution based on an adoption of LMs is proposed. Finally, to sharpen the extracted roads, a CRF method is added to our framework. The experiments were conducted on Massachusetts road aerial imagery as well as THEOS satellite imagery data sets. The results showed that our proposed framework outperformed Segnet, the state-of-the-art object segmentation technique on any kinds of remote sensing imagery, in most of the cases in terms of precision, recall, and F1.

This study employs the reflection symmetry decomposition (RSD) method to extract polarization scattering features from ground object images, aiming to determine the optimal data input scheme for deep learning networks in polarimetric synthetic aperture radar classification. Eight distinct polarizing feature combinations are designed, and the classification accuracy of various approaches is evaluated using the classic convolutional neural networks(CNNs) AlexNet and VGG16. The findings reveal that the commonly employed 6-parameter input scheme, favored by many researchers, lacks the comprehensive utilization of polarization information and warrants attention. Intriguingly, leveraging the complete 9-parameter input scheme based on the polarization coherence matrix results in improved classification accuracy. Furthermore, the input scheme incorporating all 21 parameters from the RSD and polarization coherence matrix notably enhances overall accuracy and the Kappa coefficient compared to the other 7 schemes. This comprehensive approach maximizes the utilization of polarization scattering information from ground objects, emerging as the most effective CNN input data scheme in this study. Additionally, the classification performance using the second and third component total power values (P2 and P3) from the RSD surpasses the approach utilizing surface scattering power value (PS) and secondary scattering power value (PD) from the same decomposition.

With the rapid expansion of intelligent resource-constrained devices and high-speed communication technologies, Internet of Things (IoT) has earned a wide recognition as the primary standard for low-power lossy networks (LLNs). Nevertheless, IoT infrastructures are vulnerable to cyber-attacks due to the constraints in computation, storage, and communication capacity of the endpoint devices. From one side, the majority of newly developed cyber-attacks are formed by slightly mutating formerly established cyber-attacks to produce a new attack tending to be treated as a normal traffic through the IoT network. From the other side, the influence of coupling the deep learning techniques with cybersecurity field has become a recent inclination of many security applications due to their impressive performance. In this paper, we provide a comprehensive development of a new intelligent and autonomous deep learning-based detection and classification system for cyber-attacks in IoT communication networks leveraging the power of convolutional neural networks, abbreviated as (IoT-IDCS-CNN). The proposed IoT-IDCS-CNN makes use of the high-performance computing employing the robust CUDA based Nvidia GPUs and the parallel processing employing the high-speed I9-Cores based Intel CPUs. In particular, the proposed system is composed of three subsystems: Feature Engineering subsystem, Feature Learning subsystem and Traffic classification subsystem. All subsystems are developed, verified, integrated, and validated in this research. To evaluate the developed system, we employed the NSL-KDD dataset which includes all the key attacks in the IoT computing. The simulation results demonstrated more than 99.3% and 98.2% of cyber-attacks’ classification accuracy for the binary-class classifier (normal vs anomaly) and the multi-class classifier (five categories) respectively. The proposed system was validated using k-fold cross validation method and was evaluated using the confusion matrix parameters (i.e., TN, TP, FN, FP) along with other classification performance metrics including precision, recall, F1-score, and false alarm rate. The test and evaluation results of the IoT-IDCS-CNN system outperformed many recent machine-learning based IDCS systems in the same area of study.

In the current natural environment, due to the complexity of the background and the high similarity of the color between immature green tomato and plant, the occlusion of the key organs (flower and fruit) by the leaves and stems will lead to low recognition rate and poor generalization of the detection model. Therefore, an improved tomato organ detection method based on convolutional neural network has been proposed in this paper. Based on the original Faster R-CNN algorithm, Resnet-50 with residual blocks was used to replace the traditional vgg16 feature extraction network, and K-means clustering method was used to adjust more appropriate anchor size than manual setting to improve detection accuracy. A variety of data augmentation techniques were used to train the network. The test results showed that compared with the traditional Faster R-CNN model, the mean average precision (mAP) of the optimal model was improved from 85.2% to 90.7%, the memory requirement decreased from 546.9MB to 115.9 MB, and the average detection time was shortened to 0.073S/sheet. As the performance greatly improved, the training model can be transplanted to the embedded system, which lays a theoretical foundation for the development of precise targeting pesticide application system and automatic picking device.

The timely diagnosis of defects at their incipient stage of formation is crucial to extend the life-cycle of technical appliances. This is the case of mechanical-related stress, either due to long aging degradation processes (e.g., corrosion) or in-operation forces (e.g., impact events), which might provoke detrimental damages, such as cracks, disbonding or delaminations, most commonly followed by the release of acoustic energy. The localization of these sources can be successfully fulfilled via adoption of Acoustic Emission (AE)-based inspection techniques through the computation of the Time of Arrival (ToA), namely the time at which the induced mechanical wave released at the occurrence of the acoustic event arrives to the acquisition unit. However, the accurate estimation of the ToA may be hampered by poor Signal-to-Noise ratios (SNRs). In these conditions, standard statistical methods typically fail. In this work, two alternative Deep Learning methods are proposed for ToA retrieval, namely a Dilated Convolutional Neural Network (DilCNN) and a Capsule Neural Network for ToA (CapsToA). These methods have the additional benefit of being portable on resource-constrained microprocessors. Their performance has been extensively studied on both synthetic and experimental data, focusing on the problem of ToA identification for the case of a metallic plate. Results show that the two novel methods can achieve localization errors which are up to 70% more precise than those yielded by conventional strategies, even when the SNR is severely compromised (i.e., down to 2 dB). Moreover, DilCNN and CapsNet have been implemented in a tiny machine learning environment and then deployed on microcontroller units, showing a negligible loss of performance with respect to offline realizations.

This research investigates the application of deep learning in sentiment analysis of Canadian maritime case law. It offers a framework for improving maritime law and legal analytics policy-making procedures. The automation of legal document extraction takes center stage, underscoring the vital role sentiment analysis plays at the document level. Therefore, this study introduces a novel strategy for sentiment analysis in Canadian maritime case law, combining sentiment case law approaches with state-of-the-art deep learning techniques. The overarching goal is to systematically unearth hidden biases within case law and investigate their impact on legal outcomes. Employing Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) based models, this research achieves a remarkable accuracy of 98.05% for categorizing instances. In contrast, Conventional Machine Learning techniques such as Support Vector Machines (SVM) yield an accuracy rate of 52.57%, Naive Bayes at 57.44 %, and Logistic Regression at 61.86%. The superior accuracy of the CNN and STM model combination underscores its usefulness in legal sentiment analysis, offering promising future applications in diverse fields like legal analytics and policy design. These findings mark a significant choice for AI-powered legal tools, presenting more sophisticated and sentiment-aware options for the legal profession.

Machine learning (ML) is enabling augmented reality (AR) to gain popularity in various fields, including gaming, entertainment, healthcare, and education. ML enhances AR applications in education by providing accurate visualizations of objects. For AR systems, ML algorithms facilitate the recognition of objects and gestures from kindergarten through university. The purpose of this survey is to provide an overview of various ways in which ML techniques can be applied within the field of AR within education. The first step is to describe the background of AR. In the next step, we will discuss the ML models that are used in AR education applications. Additionally, we discuss how ML is used in AR. Each subgroup’s challenges and solutions can be identified by analyzing these frameworks. In addition, we outline several research gaps and future research directions in ML-based AR frameworks for education.

Despite numerous studies in statistical downscaling methodology, there remains a lack of methods that can downscale from precipitation modeled in global climate models to regional level high resolution gridded precipitation. This paper reports a novel downscaling method using a Generative Adversarial Network (GAN), CliGAN, which can downscale large-scale annual maximum precipitation given by simulation of multiple atmosphere-ocean global climate models (AOGCM) from Coupled Model Inter-comparison Project 6 (CMIP6) to regional-level gridded annual maximum precipitation data. This framework utilizes a convolution encoder-dense decoder network to create a generative network and a similar network to create a critic network. The model is trained using an adversarial training approach. The critic uses the Wasserstein distance loss function and the generator is trained using a combination of adversarial loss Wasserstein distance, structural loss with the multi-scale structural similarity index (MSSIM), and content loss with the Nash-Sutcliff Model Efficiency (NS). The MSSIM index allowed us to gain insight into the model’s regional characteristics and shows that relying exclusively on point-based error functions, widely used in statistical downscaling, may not be enough to reliably simulate regional precipitation characteristics. Further use of structural loss functions within CNN-based downscaling methods may lead to higher quality downscaled climate model products.

Speech emotion recognition (SER) is a technology that can be applied in distance education to analyze speech patterns and evaluate speakers’ emotional states in real-time. It provides valuable insights and can be used to enhance the learning experience by enabling the assessment of instructors’ emotional stability, a factor that significantly impacts information delivery effectiveness. Students demonstrate different engagement levels during learning activities, and assessing this engagement is an important aspect of controlling the learning process and improving e-learning systems. An important aspect that may influence student engagement is the emotional states of their instructors. Accordingly, this research uses deep learning techniques to create an automated system for recognizing instructors’ emotions in their speech when delivering distance learning. This methodology entails integrating Transformer, convolutional neural network, and long short-term memory architectures into an ensemble to enhance SER. Feature extraction from audio data used Mel-frequency cepstral coefficients, chroma, Mel spectrogram, zero crossing rate, spectral contrast, centroid, bandwidth, roll-off, and root-mean square, with subsequent optimization processes adding noise to, conducting time stretching, and shifting the audio data. Notably, several Transformer blocks were incorporated, and a multi-head self-attention mechanism was employed to identify the relationships between the input sequence segments. The pre-processing and data augmentation methodologies significantly enhanced the precision of the results in that the model achieved accuracy rates of 96.3%, 99.86%, 96.5%, and 85.3% on the Ryerson Audio-Visual Database of Emotional Speech and Song, Berlin Database of Emotional Speech, Surrey Audio-Visual Expressed Emotion, and Interactive Emotional Dyadic Motion Capture datasets. Furthermore, it achieved 83% accuracy on another dataset created for this research—the Saudi Higher Education Instructor Emotions dataset. The results demonstrate this model’s considerable accuracy in detecting emotions in speech data across different languages and datasets.

Macula fovea detection is a crucial prerequisite towards screening and diagnosing macular diseases. Without early detection and proper treatment, any abnormality involving the macula may lead to blindness. However, with the ophthalmologist shortage and time-consuming artificial evaluation, neither accuracy nor effectiveness of the diagnose process could be guaranteed. In this project, we proposed a deep learning approach on ultra-widefield fundus (UWF) images for macula fovea detection. This study collected 2300 ultra-widefield fundus images from Shenzhen Aier Eye Hospital in China. Methods based on U-shape network (Unet) and Fully Convolutional Networks (FCN) are implemented on 1800 (before amplifying process) training fundus images, 400 (before amplifying process) validation images and 100 test images. Three professional ophthalmologists were invited to mark the fovea. A method from the anatomy perspective is investigated. This approach is derived from the spatial relationship between macula fovea and optic disc center in UWF. A set of parameters of this method is set based on the experience of ophthalmologists and verified to be effective. Results are measured by calculating the Euclidean distance between proposed approaches and the accurate grounded standard, which is detected by Ultra-widefield swept-source optical coherence tomograph (UWF-OCT) approach. Through a comparation of proposed methods, we conclude that, deep learning approach of Unet outperformed other methods on macula fovea detection tasks, by which outcomes obtained are comparable to grounded standard method.

Intelligent fault diagnosis techniques have replaced the time-consuming and unreliable human analysis, increasing the efficiency of fault diagnosis. Deep learning model can improve the accuracy of intelligent fault diagnosis with the help of its multilayer nonlinear mapping ability. This paper has proposed a novel method named Deep Convolutional Neural Networks with Wide First-layer Kernels (WDCNN). The proposed method uses raw vibration signals as input (data augmentation is used to generate more inputs), and uses the wide kernels in first convolutional layer for extracting feature and suppressing high frequency noise. Small convolutional kernels in the preceding layers are used for multilayer nonlinear mapping. AdaBN is implemented to improve the domain adaptation ability of the model. The proposed model addresses the problem that currently, the accuracy of CNN applied to fault diagnosis is not very high. WDCNN can not only achieve 100% classification accuracy on normal signals, but also outperform state of the art DNN model which is based on frequency features under different working load and noisy environment.

Background: Processing whole-slide images (WSI) to train neural networks can be intricate and laborious. We developed an open-source library covering recurrent tasks in processing of WSI and in evaluating the performance of the trained networks for classification tasks. Methods: Two histopathology use-cases were selected. First we aimed to train a CNN to distinguish H&E-stained slides obtained from neuropathologically classified low-grade epilepsy-associated dysembryoplastic neuroepithelial tumor (DNET) and ganglioglioma (GG). The second project we trained a convolutional neural network (CNN) to predict the hormone expression of pituitary adenoms only from hematoxylin and eosin (H&E) stained slides. In the same approach, we addressed the issue to also predict clinically silent corticotroph adenoma. We included four clinico-pathological disease conditions in a multilabel approach. Results: Our best performing CNN achieved an area under the curve (AUC) of 0.97 for the receiver operating characteristic (ROC) for corticotroph adenoma, 0.86 for silent corticotroph adenoma and 0.98 for gonadotroph adenoma. Our DNET-GG classifier achieved an AUC of 1.00 for the ROC curve. All scores were calculated with the help of our library on predictions on a case basis. Conclusions: Our comprehensive library is most helpful to standardize the work-flow and minimize the work-burden in training CNN. It is also compatible with fastai. Indeed, our new CNNs reliably extracted neuropathologically relevant information from the H&E staining only. This approach will supplement the clinico-pathological diagnosis of brain tumors, which is currently based on cost-intense microscopic examination and variable panels of immunohistochemical stainings.

In recent years, data providers are generating and streaming a large number of images. More particularly, processing images that contain faces have received great attention due to its numerous applications, such as entertainment and social media apps. The enormous amount of images shared on these applications presents serious challenges and requires massive computing resources to ensure efficient data processing. However, images are subject to a wide range of distortions in real application scenarios during the processing, transmission, sharing, or combination of many factors. So, there is a need to guarantee acceptable delivery content, even though some distorted images do not have access to their original version. In this paper, we present a framework developed to estimate the images' quality while processing a large number of images in real-time. Our quality evaluation is measured using an integration of a deep network with random forests. In addition, a face alignment metric is used to assess the facial features. Experimental results have been conducted on two artificially distorted benchmark datasets, LIVE and TID2013. We show that our proposed approach outperforms the state-of-art methods, having a Pearson Correlation Coefficient (PCC) and Spearman Rank Order Correlation Correlation Coefficient (SROCC) with subjective human scores of almost 0.942 and 0.931 while minimizing the processing time from 4.8ms to 1.8ms.

Electricity load forecasting is a crucial undertaking within all the deregulated markets globally. In contemporary times, the transition from conventional electricity grids to Smart Grids constitutes an area where extensive research is conducted on a global scale. Among the research challenges, the investigation of Deep Transfer Learning (DTL) in the field of electricity load forecasting represents a fundamental effort that imparts generality to Artificial Intelligence applications, due to new capabilities, such as knowledge transfer and computational power reduction. In this paper a comprehensive study is conducted for day-ahead electricity load forecasting. For this purpose, three Sequence-to-Sequence (Seq2seq) Deep Learning (DL) models are used, namely the Multilayer Perceptron (MLP), the Convolutional Neural Network (CNN) and the Ensemble Learning Model (ELM), which is consisted of the weighted combination of the outputs of MLP and CNN models. Also, the study focuses on the development of different forecasting strategies based on DTL and emphasizing the way the datasets are trained and fine-tuning for higher forecasting accuracy. In order to implement the forecasting strategies using Deep Learning models, load datasets from three Greek islands, Rhodes, Lesvos, and Chios, are used. The main purpose is to apply DTL for day-ahead predictions (1-24 hours) for each month of the year for Chios dataset after training and fine-tuning the models using the datasets of the three islands in various combinations. After several trials, four DTL strategies are illustrated. In the first strategy (DTL Case 1), each of the three DL models is trained using only Lesvos dataset, while fine-tuning is performed on the dataset of Chios island, in order to create day-ahead predictions for Chios load. In the second strategy (DTL Case 2), data from both Lesvos and Rhodes concurrently are used for the DL model training period, and fine-tuning is performed on the data from Chios. The third DTL strategy (DTL Case 3) contains the training of the DL models using Lesvos dataset, and the testing period performed directly on the Chios dataset without fine-tuning. The fourth strategy is a Multi-task Deep Learning (MTDL) approach, which has been extensively studied in recent years. In MTDL, the three DL models are trained simultaneously on all three datasets and the final predictions are made on the unknown part of the dataset of Chios. In this paper, we explore the performance of DTL and compare the results with those produced with MTDL. The results demonstrated that DTL can be applied with high efficiency for day-ahead load forecasting. Specifically, the two cases with fine-tuning (DTL Case 1 and 2) outperformed MTDL in terms of load prediction accuracy. Regarding the DL models, all three exhibit very high prediction accuracy, especially in the two cases with fine-tuning. The ELM excels compared to the single models. More specifically, for conducting day-ahead predictions, it has been concluded that the MLP model presents best monthly forecasts with a MAPE of 6.24% and 6.01% for the first two cases, the CNN model presents best monthly forecasts with a MAPE of 5.57% and 5.60% respectively and the ELM model achieves best monthly forecasts with a MAPE of 5.29% and 5.31%, respectively, indicating the very high accuracy it can achieve.

Amidst the burgeoning global concerns surrounding skin cancer, the quest for precise and efficient diagnostic methodologies has intensified. Conventional diagnostic approaches relying on subjective visual inspection by dermatologists often grapple with subjectivity and resource constraints. This predicament underscores the critical role of Artificial Intelligence (AI) in revolutionizing diagnostic paradigms by significantly enhancing accuracy and accessibility. This study embarks on an extensive exploration of skin cancer classification, specifically targeting the intricate task of 8-class skin cancer classification. Employing cutting-edge deep learning models on the ISIC2019 dataset, a comprehensive analysis was conducted utilizing a diverse array of pre-trained ImageNet architectures and Vision Transformer models. To counteract the inherent class imbalance in skin cancer datasets, a pioneering "Naturalize" augmentation technique was introduced, leading to the creation of indispensable tools—the Naturalized 2.4K ISIC2019 and groundbreaking Naturalized 7.2K ISIC2019 datasets—catalyzing advancements in classification accuracy. The pivotal role of AI in mitigating the risks of misdiagnosis and under-diagnosis in skin cancer becomes evident through this research. AI’s proficiency in analyzing vast datasets and discerning subtle patterns augments the diagnostic prowess of dermatologists significantly. Integration of AI-powered systems in skin cancer diagnosis holds immense promise in facilitating early detection and substantially improving patient outcomes. The meticulous evaluation employed quantitative measures such as confusion matrices, classification reports, and visual analyses using Score-CAM across diverse dataset variations. Astonishingly, the culmination of these endeavors culminated in an unprecedented achievement—100% average accuracy, precision, recall, and F1-Score—within the groundbreaking Naturalized 7.2K ISIC2019 dataset. This remarkable feat underscores the transformative potential of AI-driven methodologies, exemplifying their prowess in redefining the landscape of skin cancer diagnosis and management. In summary, this research represents a pivotal stride towards redefining dermatological diagnosis, showcasing the remarkable impact of AI-powered solutions in surmounting the challenges inherent in skin cancer diagnosis. The attainment of 100% across crucial metrics within the Naturalized 7.2K ISIC2019 dataset serves as a testament to the transformative capabilities of AI-driven approaches in reshaping the trajectory of skin cancer diagnosis and patient care.

Accurate forecasting of wind power generation plays a key role in improving the operation and management of a power system network and thereby its reliability and security. However, predicting wind power is complex due to the existence of high non-linearity in wind speed that eventually relies on prevailing weather conditions. In this paper, a novel hybrid deep learning model is proposed to improve the prediction accuracy of very short-term wind power generation for the Bodangora Wind Farm located in New South Wales, Australia. The hybrid model consists of convolutional layers, gated recurrent unit (GRU) layers and a fully connected neural network. The convolutional layers have the ability to automatically learn complex features from raw data while the GRU layers are capable of directly learning multiple parallel sequences of input data. The data sets of five-minute intervals from the wind farm are used in case studies to demonstrate the effectiveness of the proposed model against other advanced existing models, including long short-term memory (LSTM), GRU, autoregressive integrated moving average (ARIMA) and support vector machine (SVM), which are tuned to optimise outcome. It is observed that the hybrid deep learning model exhibits superior performance over other forecasting models to improve the accuracy of wind power forecasting, numerically, up to 1.59 per cent in mean absolute error, 3.73 per cent in root mean square error and 8.13 per cent in mean absolute percentage error.

Microarray gene expression-based detection and classification of medical conditions have been prominent in research studies over the past few decades. However, extracting relevant data from the high-volume microarray gene expression with inherent nonlinearity and inseparable noise components raises significant challenges during data classification and disease detection. So, this paper proposes a two-level strategy involving feature extraction and selection methods before the classification step. The feature extraction step utilizes Short Term Fourier Transform (STFT), and the feature selection step employs Particle Swarm Optimization (PSO) and Harmonic Search (HS) metaheuristic methods. The classifiers employed are Non-Linear Regression, Gaussian Mixture Model, Softmax Discriminant, Naive Bayes, SVM (Linear), SVM (Polynomial), and SVM (RBF). The two-level extracted relevant features are compared with raw data classification results, including Convolutional Neural Network (CNN) Methodology. Among the methods, STFT with PSO feature selection and SVM (RBF) classifier produced the highest accuracy of 94.47%.

This work addresses the challenge of EEG visual multiclass classification into 40 classes for Brain-Computer interface applications, by using deep learning architectures. The visual multiclass classification approach offers BCI applications a significant advantage, since it allows the supervision of more than one BCI interaction, considering that each class label supervises a BCI task. However, because of the nonlinearity and nonstationarity of EEG signals, performing multiclassifiers based on EEG features remains a significant challenge for BCI systems. In the present work, mutual information-based discriminant channel selection and Minimum-Norm Estimate algorithms are implemented to select discriminant channels and enhance EEG data. Hence, deep EEGNet and Convolutional-recurrent neural networks are implemented separately to classify EEG data of image visualization into 40 labels. By using the k-fold cross-validation approach, average classification accuracies of 94.8% and 89.8% were obtained by implementing the aforementioned network architectures, respectively. Satisfactory results obtained with this method offer a new implementation opportunity for multi-task BCI applications by utilizing a reduced number of channels (<50%), compared to those presented in the related literature where the whole set of channels is used.

Speech recognition, also known as automatic speech recognition (ASR), is a technology that enables software to transcribe spoken language into text. However, existing Amharic ASR methods require multiple separate blocks, such as language, acoustic, and pronunciation models with dictionaries, which can be time-consuming and influence performance. This study proposes an approach that replaces much of the speech pipeline with a single recurrent neural network (RNN) architecture. Our proposed architecture is based on a hybrid approach that combines a convolutional neural network (CNN) with a recurrent neural network (RNN) and a connectionist temporal classification (CTC) loss function. We conducted several experiments with noisy audio data that contain 20,000 valid sentences. The model was evaluated using the word error rate (WER) metric, achieving impressive results of 7% WER on noisy data. This approach has significant implications for the field of speech recognition, as it reduces the human effort required to create dictionaries and improves the efficiency and accuracy of ASR systems, making them more practical for real-world applications.

Drowsiness detection is an important task in road safety and other areas that require sustained attention. In this article, an approach to detect drowsiness in drivers is presented, focusing on the eye region, since eye fatigue is one of the first symptoms of drowsiness. The method used for the extraction of the eye region is Mediapipe, chosen for its high accuracy and robustness. Three neural networks were analyzed based on: InceptionV3, VGG16 and ResNet50V2 that implement deep learning. The database used is NITYMED, which contains videos of drivers with different levels of drowsiness. The three networks were evaluated in terms of accuracy, precision and recall in detecting drowsiness in the eye region. The results of the study show that all three convolutional neural networks have high accuracy in detecting drowsiness in the eye region. In particular, the Resnet50V2 network achieved the highest accuracy, with a rate of 99.71% on average. For better visualization of the data, the Grad-CAM technique is used, with which we get a better understanding of the performance of the algorithms in the classification process.

One of the critical tools for early detection and subsequent evaluation of the incidence of lung diseases is chest radiography. This study presents a real-world implementation of a convolutional neural network (CNN) based Carebot Covid app to detect COVID-19 from chest X-ray (CXR) images. Our proposed model takes the form of a simple and intuitive application. Used CNN can be deployed as a STOW-RS prediction endpoint for direct implementation into DICOM viewers. The results of this study show that the deep learning model based on DenseNet and ResNet architecture can detect SARS-CoV-2 from CXR images with precision of 0.981, recall of 0.962 and AP of 0.993.

Deep Learning (DL) oriented document processing is widely used in different fields for extraction, recognition, and classification processes from raw corpora of data. The article examines the application of deep learning approaches, based on different neural network methods, including Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), and Convolutional Neural Networks (CNN). The compared models were combined with two different word embedding techniques, namely: Bidirectional Encoder Representations from Transformers BERT and Gensim Word2Vec. The models are designed to evaluate the performance of architectures based on neural network techniques for the classification of CVs of Moroccan engineering students at ENSAK(National School of Applied Sciences of Kenitra, Ibn Tofail University). The used dataset included resumes collected from engineering students at ENSAK in 2023 for a project on the employability of Moroccan engineers in which new approaches were applied, especially machine learning, deep learning, and big data. Accordingly, 867 resumes were collected from five specialties of study (Electrical Engineering, Networks and Systems Telecommunications, Computer Engineering, Automotive Mechatronics Engineering, Industrial Engineering). The results revealed good performance of the proposed models based on the BERT embedding approach compared to models based on the Gensim Word2Vec embedding approach. Accordingly, the CNN-GRU/Bert model achieved slightly better accuracy with 0.9251 compared to other hybrid models.

An effective process monitoring strategy is a requirement for meeting the challenges posed by increasingly complex products and manufacturing processes. To address these needs, this study investigates a comprehensive scheme based on classical machine learning methods, deep learning algorithms, and feature extraction and selection techniques. In a first step, a novel deep learning architecture based on convolutional neural networks (CNN) and gated recurrent units (GRU) is introduced to predict the local weld quality based on mid-wave infrared (MWIR) and near-infrared (NIR) image data. The developed technology is used to discover critical welding defects including lack of fusion (false friends), sagging and lack of penetration, and geometric deviations of the weld seam. Additional work is conducted to investigate the significance of various geometrical, statistical, and spatio-temporal features extracted from the keyhole and weld pool regions. Furthermore, the performance of the proposed deep learning architecture is compared to that of classical supervised machine learning algorithms, such as multi-layer perceptron (MLP), logistic regression (LogReg), support vector machines (SVM), decision trees (DT), random forest (RF) and k-Nearest Neighbors (kNN). Optimal hyperparameters for each algorithm are determined by an extensive grid search. Ultimately, the three best classification models are combined into an ensemble classifier that yields the highest detection rates and achieves the most robust estimation of welding defects among all classifiers studied, which is validated on previously unknown welding trials.

Advancements in technology have revolutionized communication on social media platforms, where reviews and comments greatly influence product selection. Current opinion mining methods predominantly focus on textual content, overlooking the rich information within customer-posted images, termed here as Multus-Medium. This research introduces an innovative deep learning approach, Multus Medium Opinion Mining (MMOM), capitalizing on both text and image data for comprehensive product review analysis. MMOM employs an integrated model of Bidirectional Long Short-Term Memory (BiLSTM) and embedded Convolutional Neural Network (CNN), incorporating architectures of GoogleNet and VGGNet, thus enabling efficient extraction and fusion of textual and visual features. This synergistic approach enables data collection, preprocessing, feature extraction, fusion strategy-based feature vector generation, and subsequent product recommendation. Performance evaluation on two diverse real-world datasets name as “flicker8k” and “t4sa”. These datasets show substantial improvement over existing methods. MMOM outperforms standard benchmark models, achieving an accuracy, F1 score and ROC over fliker8k are 90.38%, 88.75% and 93.08%, whereas for twitter dataset are 88.54%, 86.34%, and 92.26% respectively, the accuracy of purposed model is 7.34% and 9.54% higher than the other two mentioned techniques. These statistics highlight the robustness and applicability of MMOM across various domains. The compelling results underscore the potential of MMOM, providing a more holistic and precise approach to opinion mining in the era of social media product reviews. This research, product recommendation helps the customer to make purchasing decision. Last but not the least, the purposed scheme can further be expanded in any other sentiment task like hospital recommendation system, crop farming recommendation and medical diagnostic system etc.

Alzheimer’s Disease (AD) affects the quality of life of millions of people worldwide and represents one of the biggest challenges for the whole society. The SH-SY5Y neuroblastoma cell line is often used as an in vitro model of neuronal function and is widely applied to study the molecular events leading to AD. In the last few years, basic research on SH-SY5Y cells has provided interesting insights for the discovery of new drugs and biomarkers for improved AD treatment and diagnosis. At the same time, untargeted NMR metabolomics is widely applied on biological fluids for (i) metabolic profile analysis, (ii) screening for differential metabolites, (iii) analysis of metabolic pathways, and (iv) the discovery of new biomarkers. Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) have proved to be powerful methods for processing NMR data, being useful in signal quantization, even if more sophisticated --- typically non--linear --- techniques are needed to obtain compact yet information--rich embeddings for complex spectra. In this paper, a compression technique based on convolutional autoencoders is proposed, which can perform a high dimensionality reduction of the spectral signal (up to more than 300 times), maintaining informative features (guaranteed by a reconstruction error always smaller than 5%). Moreover, before compression, an ad hoc preprocessing method was devised to remedy the scarcity of available data. The compressed spectral data were then used to train some SVM classifiers to distinguish diseased from healthy cells, achieving an accuracy close to 78%, a significantly better performance with respect to using PCA--compressed data.

In recent times, with the increase of Artificial Neural Network (ANN), deep learning has brought a dramatic twist in the field of machine learning by making it more Artificial Intelligence (AI). Deep learning is used remarkably used in vast ranges of fields because of its diverse range of applications such as surveillance, health, medicine, sports, robotics, drones etc. In deep learning, Convolutional Neural Network (CNN) is at the center of spectacular advances that mixes Artificial Neural Network (ANN) and up to date deep learning strategies. It has been used broadly in pattern recognition, sentence classification, speech recognition, face recognition, text categorization, document analysis, scene, and handwritten digit recognition. The goal of this paper is to observe the variation of accuracies of CNN to classify handwritten digits using various numbers of hidden layer and epochs and to make the comparison between the accuracies. For this performance evaluation of CNN, we performed our experiment using Modified National Institute of Standards and Technology (MNIST) dataset. Further, the network is trained using stochastic gradient descent and the backpropagation algorithm.

In computer vision and image analysis, Convolutional Neural Networks (CNNs) and other deep learning models are at the forefront of research and development. These advanced models have proven to be highly effective in tasks related to computer vision. One technique that has gained prominence in recent years is the construction of ensembles using Deep CNNs. These ensembles typically involve combining multiple pre-trained CNNs to create a more powerful and robust network. The purpose of this study is to evaluate the effectiveness of building CNN ensembles by combining several advanced techniques. Tested here are CNN ensembles constructed by replacing ReLU layers with different activation functions, employing various data augmentation techniques, and utilizing several algorithms, including some novel ones, that perturb network weights. Experimental results performed across many data sets representing different tasks demonstrate that our proposed methods for building deep ensembles produces superior results. All the resources required to replicate our experiments are available at https://github.com/LorisNanni.

Stress has emerged as a major concern in modern society, significantly impacting human health and well-being. Statistical evidence underscores the extensive social influence of stress, especially in terms of work-related stress and associated healthcare costs. This work addresses the critical need for accurate stress detection, emphasising its far-reaching effects on health and social dynamics. Focusing on remote stress monitoring, this work proposes an efficient deep learning framework to discern stress from facial videos. In contrast to research on wearable devices, this paper investigates the application of *dl for stress detection based on *rppg. The methodology involves selecting suitable *dl models (*lstm, *gru, *1dcnn), optimising hyperparameters, and investigating augmentation techniques. 1D-CNNv1 model achieved the best performance compared to other approaches, particularly with augmentation techniques such as linear interpolation and white noise, achieving a stress classification accuracy of 95.83% while maintaining excellent computational efficiency. The experimental results demonstrate the use of *dl for stress detection based on *rppg, with the potential to make significant contributions to the international standard in the field.

The prompt and accurate diagnosis of breast lesions, including the distinction between cancer, non-cancer, and suspicious cancer, plays a crucial role in the prognosis of breast cancer. In this paper, we introduce a novel method based on feature extraction and reduction for detection of breast cancer in mammography images. First, we extract features from multiple pre-trained convolutional neural network (CNN) models, and then concatenate them. The most informative features are selected based on their mutual information with the target variable. Subsequently, the selected features can be classified using a machine learning algorithm. We evaluate our approach using four different machine learning algorithms, and our results demonstrate that the neural network-based classifier yields an accuracy as high as 92% for the RSNA dataset which is a new dataset that provides two views and additional features such as age. We compare our proposed algorithm with state-of-the-art methods and demonstrate its superiority, particularly in terms of accuracy and sensitivity. For the MIAS dataset, we achieve an accuracy as high as 94.5%, and for the DDSM dataset, an accuracy of 96% is attained. These results highlight the effectiveness of our method in accurately diagnosing breast lesions and surpassing existing approaches.

Reliable prediction of infectious disease is an essential need for public health organizations to minimize or prevent the disease's spread. With the increase of data growth in the healthcare sector, accurate analysis of such data helps for early disease detection and better patient care. With the availability of huge data, it is a great deal for predicting and handling an epidemic occurrence. To control the epidemic disease, a new Artificial Intelligence called Manhattan Hessian Linear Feature Embedding Based Tucker Congruence Deep Convolutional Network (MHLFE-TCDCN) is introduced with higher accuracy and minimum time consumption. The MIoT device is used to collect the data from the patient body. The proposed MHLFE-TCDCN includes two major processes namely feature selection and classification with multiple layers such as input, two hidden layers, and one output layer. First, the feature selection process is performed using the Manhattan Hessian Locally Linear Embedding technique in the first hidden layer to minimize the dimensionality and time complexity. After selecting the feature, the classification is performed using the Tucker coefficient of congruence regression. The Tucker coefficient of congruence measures the similarity between the training and testing disease data. A positive similarity level is used to accurately identify the disease at the output layer based on the similarity measure. Experimental assessment is carried out with various parameters such as accuracy, precision, recall, F-measure, and computational time concerning several numbers of patient data. The quantitative results and discussion that the proposed MHLFE-TCDCN technique achieves higher accuracy with minimal computation time as compared to the conventional methods.

This paper presents a study and implementation of a convolutional neural network to identify and recognize humpback whale specimens from the unique patterns of their tails. Starting from a dataset composed of images of whale tails, all the phases of the process of creation and training of a neural network are detailed – from the analysis and pre-processing of images to the elaboration of predictions, using TensorFlow and Keras frameworks. Other possible alternatives are also explained when it comes to tackling this problem and the complications that have arisen during the process of developing this paper.

Chinese calligraphy, revered globally for its therapeutic and mindfulness benefits, encompasses styles such as Regular (Kai Shu), Running (Xing Shu), Official (Li Shu), and Cursive (Cao Shu) scripts. Beginners often start with Regular script, advancing to more intricate styles like Cursive. Each style, marked by unique historical calligraphers' contributions, requires learners to discern distinct nuances. The integration of AI in calligraphy analysis, collection, recognition, and classification are pivotal. This study introduces an innovative Convolutional Neural Network (CNN) architecture, pioneering the application of CNN in the classification of Chinese calligraphy. Focusing on the four principal calligraphers' styles from the Tang dynasty (690-907 A.D), this research spotlights the era when the traditional regular script font (Kai Shu) was refined. A comprehensive dataset of 8282 samples from these calligraphers, representing the zenith of regular style, was compiled for CNN training and testing. The model distinguishes personal styles for classification, showing superior performance over existing networks. Achieving 89.5-96.2% accuracy in calligraphy classification, our approach underscores the significance of CNN in both font and artistic style categorization. This research paves the way for advanced studies in Chinese calligraphy and its cultural implications.

Facial expression recognition is crucial for understanding human emotions and nonverbal communication. With the growing prevalence of facial recognition technology and its various applications, accurate and efficient facial expression recognition has become a significant research area. However, most previous methods have focused on designing unique deep-learning architectures while overlooking the loss function. This study presents a new loss function that allows simultaneous consideration of inter- and intra-class variations to be applied to CNN architecture for facial expression recognition. More concretely, this loss function reduces the intra-class variations by minimizing the distances between the deep features and their corresponding class centers. It also increases the inter-class variations by maximizing the distances between deep features and their non-corresponding class centers, and the distances between different class centers. Numerical results from several benchmark facial expression databases, such as Cohn-Kanade Plus, Oulu-Casia, MMI, and FER2013, are provided to prove the capability of the proposed loss function compared with existing ones.

Creating intelligent systems capable of recognizing emotions is a difficult task, especially when looking at emotions in animals. This paper describes the process of designing a “proof of concept” system to recognize emotions in horses. This system is formed by two elements, a detector and a model. The detector is a fast region-based convolutional neural network that detects horses in an image. The model is a convolutional neural network that predicts the emotions of those horses. These two elements were trained with multiple images of horses until they achieved high accuracy in their tasks. 400 images of horses were collected and labeled to train both the detector and the model while 80 were used to validate the system. Once the two components were validated, they were combined into a testable system that would detect equine emotions based on established behavioral ethograms indicating emotional affect through head, neck, ear, muzzle and eye position. The system showed an accuracy of between 69% and 74% on the validation set, demonstrating that it is possible to predict emotions in animals using autonomous intelligent systems. Such a system has multiple applications including further studies in the growing field of animal emotions as well as in the veterinary field to determine the physical welfare of horses or other livestock.

One of the most effective image processing techniques is the use of convolutional neural networks, where we combine intensity values at grid points in the vicinity of each point. To speed up computations, researchers have developed a dilated version of this technique, in which only some points are processed. It turns out that the most efficient case is when we select points from a sub-grid. In this paper, we explain this empirical efficiency proving that the sub-grid is indeed optimal – in some reasonable sense. To be more precise, we prove that all reasonable optimality criteria, the optimal subset of the original grid is either a sub-grid, or a sub-grid-like set.

Abstract: The surge of diabetes poses a significant global health challenge, particularly in Oman and the Middle East. Early detection of diabetes is crucial for proactive intervention and improved patient outcomes. This research leverages the power of machine learning, specifically, Convolutional Neural Networks (CNNs), to develop an innovative 4D CNN model dedicated to early diabetes prediction. A region-specific dataset from Oman is utilized to enhance health outcomes for individuals at risk of developing diabetes. The proposed model showcases remarkable accuracy, achieving an average accuracy of 98.49% to 99.17% across various epochs. Additionally, it demonstrates excellent F1 score, recall, and sensitivity, highlighting its ability to identify true positive cases. The findings contribute to the ongoing effort to combat diabetes and pave the way for future research in using deep learning for early disease detection and proactive healthcare.

With the fast development of the mobile Internet, the online platforms of social networks have rapidly been developing for the purpose of making friends, sharing information, etc. In these online platforms, users being related to each other forms social networks. Literature reviews have shown that social networks have community structure. Through the studies of community structure, the characteristics and functions of networks structure and the dynamical evolution mechanism of networks can be used for predicting user behaviours and controlling information dissemination. Therefore, this study proposes a deep community detection method which includes (1) matrix reconstruction method, (2) spatial feature extraction method and (3) community detection method. The original adjacency matrix in social network is reconstructed based on the opinion leader and nearer neighbors for obtaining spatial proximity matrix. The spatial eigenvector of reconstructed adjacency matrix can be extracted by an auto-encoder based on convolution neural network for the improvement of modularity. In experiments, four open datasets of practical social networks were selected to evaluate the proposed method, and the experimental results show that the proposed deep community detection method obtained higher modularity than other methods. Therefore, the proposed deep community detection method can effectively detect high quality communities in social networks.

Conducting computational stress-strain analysis using finite element methods (FEM) is a common approach when dealing with the complex geometries of atherosclerosis, which is a leading cause of global mortality and complex cardiovascular disease. The considerable expense linked to FEM analysis encourages the substitution of FEM with a considerably faster data-driven machine learning (ML) approach. This study investigated the potential of end-to-end deep learning tools as a more effective substitute for FEM in predicting stress-strain fields within 2D cross sections of arterial wall. We first proposed a U-Net based fully convolutional neural network (CNN) to predict the von Mises stress and strain distribution based on the spatial arrangement of calcification within arterial wall cross-sections. Further, we developed a conditional generative adversarial network (cGAN) to enhance, particularly from the perceptual perspective, the prediction accuracy of stress and strain field maps for arterial walls with various calcification quantities and spatial configurations. On top of U-Net and cGAN, we also proposed their ensemble approaches to further improve the prediction accuracy of field maps. Our dataset, consisting of input and output images, was generated by implementing boundary conditions and extracting stress-strain field maps. The trained U-Net models can accurately predict von Mises stress and strain fields, with structural similarity index scores (SSIM) of 0.854 and 0.830 and mean squared errors of 0.017 and 0.018 for stress and strain, respectively, on a reserved test set. Meanwhile, the cGAN models in a combination of ensemble and transfer learning techniques demonstrate high accuracy in predicting von Mises stress and strain fields, as evidenced by SSIM scores of 0.890 for stress and 0.803 for strain. Additionally, mean squared errors of 0.008 for stress and 0.017 for strain further support the model's performance on a designated test set. Overall, this study developed a surrogate model for finite element analysis, which can accurately and efficiently predict stress-strain fields of arterial walls regardless of complex geometries and boundary conditions.

Multiply-accumulate calculations using a memristor crossbar array is an important method to realize neuromorphic computing. However, the memristor array fabrication technology is still immature, and it is difficult to fabricate large-scale arrays with high-yield, which restricts the development of memristor-based neuromorphic computing technology. Therefore, cascading small-scale arrays to achieve the neuromorphic computational ability that can be achieved by large-scale arrays, which is of great significance for promoting the application of memristor-based neuromorphic computing. To address this issue, we present a memristor-based cascaded framework with some basic computation units, several neural network processing units can be cascaded by this means to improve the processing capability of the dataset. Besides, we introduce a split method to reduce pressure of input terminal. Compared with VGGNet and GoogLeNet, the proposed cascaded framework can achieve 93.54% Fashion-MNIST accuracy under the 4.15M parameters. Extensive experiments with Ti/AlOx/TaOx/Pt we fabricated are conducted to show that the circuit simulation results can still provide a high recognition accuracy, and the recognition accuracy loss after circuit simulation can be controlled at around 0.26%.

A convolutional neural network (CNN) is sometimes understood as a black box in the sense that while it can approximate any function, studying its structure will not give us any insights into the nature of the function being approximated. In other terms, the discriminative ability does not reveal much about the latent representation of a network. This research aims to establish a framework for interpreting the CNNs by profiling them in terms of interpretable visual concepts and verifying them by means of Integrated Gradient. We also ask the question, "Do different input classes have a relationship or are they unrelated?" For instance, could there be an overlapping set of highly active neurons to identify different classes? Could there be a set of neurons that are useful for one input class whereas misleading for a different one? Intuition answers these questions positively, implying the existence of a structured set of neurons inclined to a particular class. Knowing this structure has significant values; it provides a principled way for identifying redundancies across the classes. Here the interpretability profiling has been done by evaluating the correspondence between individual hidden neurons and a set of human-understandable visual semantic concepts. We also propose an integrated gradient-based class-specific relevance mapping approach that takes the spatial position of the region of interest in the input image. Our relevance score verifies the interpretability scores in terms of neurons tuned to a particular concept/class. Further, we perform network ablation and measure the performance of the network based on our approach.

The coronavirus disease 2019 (COVID-19) is the fastest transmittable virus caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The detection of COVID-19 using artificial intelligence techniques and especially deep learning will help to detect this virus in early stages which will reflect in increasing the opportunities of fast recovery of patients worldwide. This will lead to release the pressure off the healthcare system around the world. In this research, classical data augmentation techniques along with CGAN based on a deep transfer learning model for COVID-19 detection in chest CT scan images will be presented. The limited benchmark datasets for covid-19 especially in chest CT images is the main motivation of this research. The main idea is to collect all the possible images for covid-19 that exists until the very writing of this research and use the classical data augmentations along with CGAN to generate more images to help in the detection of the COVID-19. In this study, five different deep convolutional neural network-based models (AlexNet, VGGNet16, VGGNet19, GoogleNet, and ResNet50) have been selected for the investigation to detect the coronavirus infected patient using chest CT radiographs digital images. The classical data augmentations along with CGAN improve the performance of classification in all selected deep transfer models. The Outcomes show that ResNet50 is the most appropriate deep learning model to detect the COVID-19 from limited chest CT dataset using the classical data augmentation with testing accuracy of 82.91%.

The coronavirus (covid-19) pandemic is putting healthcare systems across the world under unprecedented and increasing pressure according to the World Health Organization (WHO). With the advances in computer algorithms and especially Artificial Intelligence, the detection of this type of virus in the early stages will help in fast recovery and help in releasing the pressure off healthcare systems. In this paper, a GAN with deep transfer learning for coronavirus detection in chest x-ray images is presented. The lack of benchmark datasets for covid-19 especially in chest x-rays images is the main motivation of this research. The main idea is to collect all the possible images for covid-19 that exists until the writing of this research and use the GAN network to generate more images to help in the detection of the virus from the available x-rays images with the highest accuracy possible. The dataset used in this research was collected from different sources and it is available for researchers to download and use it. The number of images in the collected dataset is 307 images for four different types of classes. The classes are the covid-19, normal, pneumonia bacterial, and pneumonia virus. The dataset is divided into 90% for the GAN and the training and the validation phase, while 10% used in the testing phase. The GAN helps in generating more images from the original dataset to be 30 times larger than the originally collected dataset. The GAN also help in overcoming the overfitting problem and made the proposed model more robust. Three deep transfer models are selected in this research for investigation. The models are the Alexnet, Googlenet, and Restnet18. Those models are selected based on their small number of layers on their architectures, which will reflect in reducing the complexity of the models and the consumed memory and time. Using a combination of GAN and deep transfer models prove it is efficiency according to validation, testing accuracy, and performance measurements such as precision, recall, and F1 score. Three case scenarios are tested through the paper, the first scenario which includes 4 classes from the dataset, while the second scenario includes 3 classes and the third scenario includes 2 classes. All the scenarios include the covid-19 class as it is the main target of this research to be detected. In the first scenario, the Googlenet is selected to be the main deep transfer model as it achieves 80.6% in testing accuracy. In the second scenario, the Alexnet is selected to be the main deep transfer model as it achieves 85.2% in testing accuracy, while in the third scenario which includes 2 classes(covid-19, and normal), Googlenet is selected to be the main deep transfer model as it achieves 100% in testing accuracy and 99.9% in the validation accuracy. All the performance measurement strengthen the obtained results through the research. Finally, this research may be considered one of the first trails to use GAN and deep transfer models together to help in detecting coronaviruses (covid-19) within the absence of a benchmark dataset around the world, especially in x-rays chest images.

All kinds of vessels consist of dozens of complex machineries with rotating parts and electric motors that operate continuously in a harsh environment with excess temperature and humidity, vibration, fatigue and load. A breakdown or malfunction in one of these machineries can significantly impact the vessel’s operation and safety and consequently, the safety of the crew and the environment. To maintain operational efficiency and seaworthiness, the shipping industry invests substantial resources in preventive maintenance and repairs. This research presents the economic and technical benefits of predictive maintenance over to traditional preventive maintenance, and repair by replacement approaches in the maritime domain. By leveraging modern technology and Artificial Intelligence, we can analyze real-time operating conditions of machinery, enabling early detection of potential damages and allowing for effective planning of future maintenance and repair activities. In this paper, we propose and develop a Convolutional Neural Network that is fed with raw vibration measurements acquired in a laboratory environment from the ball bearings of a motor. Then, we investigate whether the proposed network can accurately detect the functional state of ball bearings and categorize any possible failures present, contributing to improved maintenance practices in the shipping industry.

Topology optimization is a pioneering design method that can provide various candidates with high mechanical properties. However, the high-resolution for the optimum structures is highly desired, normally in turn leading to computationally intractable puzzle, especially for the famous Solid Isotropic Material with Penalization (SIMP) method. In this paper, an efficient and high-resolution topology optimization method is proposed based on the Super-Resolution Convolutional Neural Network (SRCNN) technique in the framework of SIMP. The SRCNN includes four processes, i.e. refining, path extraction & representation, non-linear mapping, and reconstruction. The high computational efficiency is achieved by a pooling strategy, which can balance the number of finite element analysis (FEA) and the output mesh in optimization process. To further reduce the high computational cost of 3D topology optimization problems, a combined treatment method using 2D SRCNN is built as another speeding-up strategy. A number of typical examples justify that the high-resolution topology optimization method adopting SRCNN has excellent applicability and high efficiency for 2D and 3D problems with arbitrary boundary conditions, any design domain shape, and varied load.

In the realm of computer vision, semantic segmentation is the task of recognizing objects in images at the pixel level. This is done by performing a classification of each pixel. The task is complex and requires sophisticated skills and knowledge about the context to identify objects’ boundaries. The importance of semantic segmentation in many domains is undisputed. In medical diagnostics, it simplifies the early detection of pathologies, thus mitigating the possible consequences. In this work, we provide a review of the literature on deep ensemble learning models for polyp segmentation and we develop new ensembles based on convolutional neural networks and transformers. The development of an effective ensemble entails ensuring diversity between its components. To this end, we combine different models (HarDNet-MSEG, Polyp-PVT, and HSNet) trained with different data augmentation techniques, optimization methods, and learning rates, which we experimentally demonstrate to be useful to form a better ensemble. Most importantly, we introduce a new method to obtain the segmentation mask which is more suitable for combining transformers in an ensemble. In our extensive experimental evaluation, the proposed ensembles exhibit state-of-the-art performance.

The sustainability of the planet and its inhabitants is in dire danger and is among the highest priorities on global agendas such as the Sustainable Development Goals (SDGs) of the United Nations (UN). Solar energy -- among other clean, renewable, and sustainable energies -- is seen as essential for environmental, social, and economic sustainability. Predicting solar energy accurately is critical to increasing reliability and stability, and reducing the risks and costs of the energy systems and markets. Researchers have come a long way in developing cutting-edge solar energy forecasting methods. However, these methods are far from optimal in terms of their accuracies, generalizability, benchmarking, and other requirements. Particularly, no single method performs well across all climates and weathers due to the large variations in meteorological data. This paper proposes SENERGY (an acronym for Sustainable Energy), a novel deep learning-based auto-selective approach and tool that, instead of generalising a specific model for all climates, predicts the best performing deep learning model for GHI forecasting in terms of forecasting error. The approach is based on carefully devised deep learning methods and feature sets through an extensive analysis of deep learning forecasting and classification methods using ten meteorological datasets from three continents. We analyse the tool in great detail through a range of metrics and methods for performance analysis, visualization, and comparison of solar forecasting methods. SENERGY outperforms existing methods in all performance metrics including Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Forecast Skills (FS), Relative Forecasting Error, and the normalised versions of these metrics. The proposed auto-selective approach can be extended to other research problems such as wind energy forecasting and predict forecasting models based on different criteria (in addition to the minimum forecasting error used in this paper) such as the energy required or speed of model execution, different input features, different optimisations of the same models, or other user preferences.

This study focuses on the impact of weak partial discharges (PDs) on defect recognition accuracy in epoxy resin through phase-resolved partial discharge (PRPD) analysis. Two measurement conditions are compared until PRPD pattern saturation: one-minute and one-hour durations. The PD data specifically target three void types located at different positions within the epoxy material. The aim is to evaluate how the presence of weak PDs at the PD extinction voltage (PDEV) influences defect recognition accuracy. This research sheds light on the potential implications of neglecting the significance of weak PD signals in defect detection. A convolutional neural network (CNN) model is trained using PRPD data recorded at the PD inception voltage (PDIV) and tested using the new PRPD data from both conditions recorded from a lower PDIV to a PDEV. The trained CNN model achieves a defect recognition accuracy of 100% for a one-hour duration, highlighting the importance of not neglecting weak PD signals. This emphasizes the significance of extended measurement duration and pattern saturation in capturing and analyzing weak PD signals for improved defect recognition. This study contributes to the advancement of practical applications by understanding the behavior of the epoxy material and enhancing defect detection techniques.

With the urgent need of precise positioning faced by internet of things (IoT) applications, the universality and cost of indoor positioning devices become key factors. Since the 5G network has been widely deployed, new opportunity is brought by tightly fusing the traditional low-cost sensors, i.e., the magnetometer. In this study, using 5G channel state information (CSI) and geomagnetic data, a multi-input convolutional neural network (CNN) localization system was proposed. First, to generate a data tensor that is easy for CNN processing, the raw data was reconstructed individually. Then, to comprehensively incorporate the features of 5G CSI and geomagnetic strength data, the ReLU function was chosen as the activation function of the convolutional layer. After that, a multi-input CNN was trained using the incorporated geomagnetic strength and CSI amplitude in the off-line side, and the trained CNN was recorded as a location fingerprint, which can be used for the user position prediction. Finally, in the online side and using a multi-input CNN, the 2-D coordinates are estimated and tested indoors in a typical conference room scenario. The results showed that longer sampling time of fingerprint data result in better uniqueness of the reference point, while the data collection time of locating points does not need to be long. Taking the positioning efficiency and accuracy into consideration, a sampling time of 3s at the reference point and 0.2s at the locating point are recommended. The positioning accuracy using the proposed method was 1.41 m, with an improvement of 22.9% compared with the 5G positioning, and an improvement of 18.0% compared with the 5G and geomagnetic fusing positioning using single CNN.

Human activities and natural events alter the earth’s surface and pose a constant threat to the environment. Thus, accurately monitoring and predicting these changes in a timely manner is crucial to provide solutions and mitigate the environmental consequences beforehand. This research introduces a novel framework, called HCD-Net, for change detection using bi-temporal hyperspectral images. The framework is based on double-stream deep feature extraction and an attention mechanism. The first stream focuses on deep feature extraction through 3D convolution layers and 3D Squeeze-and-Excitation (SE) blocks. The second stream focuses on deep feature extraction through 2D convolution and 2D SE blocks. The extracted deep features are then concatenated and fed to dense layers for decision-making. The efficiency of HCD-Net is compared to the state-of-the-art change detection methods. Additionally, the bi-temporal Airborne Visible InfraRed Imaging Spectrometer (AVIRIS) hyperspectral dataset was used for further evaluation of the change detection results. The results show that HCD-Net has a higher accuracy and the lowest false alarm rate, where the overall classification accuracy is above 96%, and the kappa coefficient is over 0.9.

In recent years, the field of deep learning achieved considerable success in pattern recognition, image segmentation and may other classification fields. There are a lot of studies and practical applications of deep learning on images, video or text classification. In this study, we suggest a method for changing the architecture of the most performing CNN models with the aim of designing new models to be used as stand-alone networks or as a component of an ensemble. We propose to replace each activation layer of a CNN (usually a ReLu layer) by a different activation function stochastically drawn from a set of activation functions: in this way the resulting CNN has a different set of activation function layers.

In order to realize the performance degradation trend prediction accurately, a prediction method based on multi-domain features and temporal convolutional network (TCN) is proposed. Firstly, construct a high-dimensional feature set in the multi-domain of vibration signals, and use comprehensive evaluation indicators to preliminarily screen performance degradation indexes with good sensitivity and strong trend. Secondly, the kernel principal component analysis (KPCA) method is adopted to eliminate redundant information between multi-domain features, and construct a health index (HI) based on convolutional auto-encoder (CAE) network. Thirdly, a TCN-based performance degradation trend prediction model is constructed, and direct multi-step prediction is used to predict the performance degradation trend of the monitored object. On this basis, the validity of the proposed method is verified using the bearing public data, and it is successfully applied to performance degradation trend prediction of rolling contact fatigue specimen. The results show that the feature set can be reduced from 14 dimensions to 4 dimensions by using KPCA, while 98.33% of the information of the original feature set is retained. Furthermore, the method of constructing HI based on CAE network is effective. The change process of the HI constructed truly reflects the performance degradation process of the rolling contact fatigue specimen. Compared with the two commonly used HI construction methods, auto-encoding (AE) network and gaussian mixture model (GMM), this method has obvious advantages. At the same time, the prediction model based on TCN can accurately predict the performance degradation of the rolling contact fatigue specimen with the root mean square error 0.0146 and the mean absolute error 0.0105, which has better performance and higher prediction accuracy than the prediction model based on the long short-term memory (LSTM) network and the gated recurrent unit (GRU). This method has general significance and may be extended to the performance degradation prediction of other mechanical equipment/parts.

Early-season crop-type data are required for a variety of agricultural monitoring and decision-making applications. The early season in this study referred to the prophase and middle of a growth season. Early-season detection of crop types remains challenging because of limited discriminative features, especially for landscapes that are characterized by complex cropping patterns. In fact, different remote-sensing satellites can increase the frequency of data acquisition, which can provide more information in the early season. Moreover, optical and radar sensors have different degrees of sensitivity to crop parameters. Therefore, the integration and application of two types of multitemporal data are of great significance to improve the accuracy and timeliness of crop type detection.In deep learning (DL), convolutional neural network (CNN) and recurrent neural network (RNN) models have great potential for temporal feature extraction. Compared with RNNs, CNNs usually have fewer parameters and are more conducive to determining early-season detection dates of different crop types, which requires a lot of training because of the need to model on different dates. Nevertheless, revisit dates and temporal intervals of different satellites are usually different, resulting in different data acquisition time series; these data cannot be directly used as input for the same convolutional layer. To address this challenge, a Dual-1DCNN was bulit based on the CNN model in this study. Moreover, an incremental training method was used to attain the network on each data acquisition date and obtained the best detection date for each crop type in the early season. A case study for Hengshui City in China was conducted using time series of Sentinel-1A (S1A) and Sentinel-2 (S2). To verify this method, classical methods support vector machine (SVM) and random forest (RF) were implemented. The results demonstrated the following: (1) the Dual-1DCNN extracted discriminative features from S1A and S2 time series at the early season by producing the highest overall accuracy (OA: 87.23%); (2) for summer maize, cotton, and common yam rhizome, the Dual-1DCNN achieved F1 values of 92.39%, 87.71%, and 84.38%, respectively, at the early season (moreover, the early seasons were almost 40, 70, and 80 days before the end of the growth seasons, respectively). These findings suggested that the Dual-1DCNN is promising for the accurate and timely detection of crop types.

The paper concludes the first research on mouth-based Emotion Recognition (ER), adopting a Transfer Learning (TL) approach. Transfer Learning results paramount for mouth-based emotion ER, because a few data sets are available, and most of them include emotional expressions simulated by actors, instead of adopting a real-world categorization. Using TL we can use fewer training data than training a whole network from scratch, thus more efficiently fine-tuning the network with emotional data and improving the convolutional neural network accuracy in the desired domain. The proposed approach aims at improving the Emotion Recognition dynamically, taking into account not only new scenarios but also modified situations with respect to the initial training phase, because the image of the mouth can be available even when the whole face is visible only in an unfavourable perspective. Typical applications include automated supervision of bedridden critical patients in an healthcare management environment, or portable applications supporting disabled users having difficulties in seeing or recognizing facial emotions. This work takes advantage from previous preliminary works on mouth-based emotion recognition using CNN deep-learning, and has the further benefit of testing and comparing a set of networks on large data sets for face-based emotion recognition well known in literature. The final result is not directly comparable with works on full-face ER, but valorizes the significance of mouth in emotion recognition, obtaining consistent performances on the visual emotion recognition domain.

Many works have recently identified the need to combine deep learning with extreme learning to strike a performance balance with accuracy especially in the domain of multimedia applications. Considering this new paradigm, namely convolutional extreme learning machine (CELM), we present a systematic review that investigates alternative deep learning architectures that use extreme learning machine (ELM) for a faster training to solve problems based on image analysis. We detail each of the architectures found in the literature, application scenarios, benchmark datasets, main results, advantages, and present the open challenges for CELM. We follow a well structured methodology and establish relevant research questions that guide our findings. We hope that the observation and classification of such works can leverage the CELM research area providing a good starting point to cope with some of the current problems in the image-based computer vision analysis.

The root system plays an irreplaceable role in plant growth. Its improvement can increase crop productivity. However, such system is still mysterious for us. The underlying mechanism has not been fully uncovered. The investigation on proteins related to the root system is an important means to complete this task. In the previous time, lack of root-related proteins makes it impossible to adopt machine learning methods for designing efficient models for the discovery of novel root-related proteins. Recently, a public database on root-related proteins was set up and machine learning methods can be applied in this field. In this study, we proposed a machine learning based model, named Graph-Root, for identification of root-related proteins. The features derived from protein sequences and one network were extracted, where the former features were processed by graph convolutional neural network and multi-head attention, and the later features abstracted the linkage between proteins. These features were fed into the fully connected layer to make prediction. The 5-fold cross-validation and independent tests suggested its good performance. It also outperformed the only one previous model, SVM-Root. Furthermore, the importance of each feature type and component in the proposed model was investigated.

Retinopathy is a human eye disease that causes changes in retinal blood vessels that leads to bleed, leak fluid and vision impairment. Symptoms of retinopathy are blurred vision, changes in color perception, red spots, and eye pain. In this paper, a new methodology based on Convolutional Neural Networks (CNN) is developed and proposed to diagnose and give a decision about the presence of retinopathy. The CNN model is trained by different images of eyes that have retinopathy and those which do not have retinopathy. The performance of the proposed model is compared with the related methods of DREAM, KNN, GD-CNN and SVM. Experimental results show that the proposed CNN performs better.

To drive safely, the driver must be aware of the surroundings, pay attention to the road traffic, and be ready to adapt to new circumstances. Most studies on driving safety focus on detecting anomalies in driver behavior and monitoring the cognitive capabilities of drivers. In our study, we proposed a classifier for basic activities in driving a car, based on a similar approach that could be applied to the recognition of basic activities in daily life, that is, using electrooculographic (EOG) signals and a one-dimensional convolutional neural network (1D CNN). Our classifier achieved an accuracy of 80% for the 16 primary and secondary activities. The accuracy related to primary activities in driving, including crossroad, parking, roundabout was 97.9%, 96.8%, 97.4%, and 99.5%, respectively. The F1 score for secondary driving actions (0.99) was higher than for primary driving activities (0.93–0.94). Furthermore, using the same algorithm, it was possible to distinguish four secondary activities related to activities of daily life and secondary when driving a car.

Convolutional Neural Networks (CNNs) are largely hand-crafted, which leads to inefficiency in the constructed network. Various other algorithms have been proposed to address this issue, but the inefficiencies resulting from human intervention have not been addressed. Our proposed EvolveNet algorithm is a task-agnostic evolutionary search algorithm that can find optimal depth and width scales automatically in an efficient way. The optimal configurations are not found using grid search, instead evolved from an existing network. This eliminates inefficiencies that emanate from hand-crafting, thus reducing the drop in accuracy. The proposed algorithm is a framework to search through a large search space of subnetworks until a suitable configuration is found. Extensive experiments on the ImageNet dataset demonstrate the superiority of the proposed method by outperforming the state-of-the-art methods.

Object detection is a popular image processing technique, widely used in numerous applications for detecting and locating objects in images or videos. While being one of the fastest algorithms for object detection, Single-Shot Multibox Detection (SSD) networks are also computationally very demanding, which limits their usage in real-time edge applications. Even though the SSD post-processing algorithm is not the most complex segment of the overall SSD object detection network, it is still computationally demanding and can become a bottleneck with respect to processing latency and power consumption, especially in edge applications with limited resources. When using hardware accelerators to accelerate backbone CNN processing, the SSD post-processing step implemented in software can become critical for high-end applications where high frame rates are required, as this paper shows. To overcome this problem, we propose Puppis, an architecture for hardware acceleration of SSD post-processing algorithm. As experiments will show, our solution will lead to an average SSD post-processing speedup of 34.42 when compared with a software implementation. Furthermore, execution of a complete SSD network will be on average 45.39 times faster than software implementation when the proposed Puppis SSD hardware accelerator is used together with some existing CNN accelerators.

In this paper, we propose methods for brain tumor detection in MRI images based on ensemble learning. We build upon prior research on ensemble methods by testing the concatenation of pre-trained models: features extracted via transfer learning are merged and segmented by classification algorithms or a stacked ensemble of those algorithms. The proposed approach achieved accuracy scores of 0.98 , outperforming a benchmark VGG-16 model. Considerations to granular computing are given in the paper as well.

This study aimed at exploring what role artificial intelligence techniques could play in the futural numerical analysis. In this paper, a convolutional neural network techniques based on modified loss function is proposed as a surrogate of finite element method(FEM). Several surrogate-based physics-informed neural networks(PINNs) are developed to solve representative boundary value problems based on elliptic partial differential equations (PDEs). Results from the proposed surrogate-based approach are in good agreement with ones from conventional FEM. It is found that modification of the loss function could improve the prediction accuracy of the neural network. It is indicated that to some extent the artificial intelligence technique could replace conventional numerical analysis as a great surrogate model.

Deep learning has recently attracted much attention due to its excellent performance in processing audio, image, and video data. However, few studies are devoted to the field of automatic modulation classification (AMC). It is one of the most well-known research topics in communication signal recognition, which remains challenging for traditional methods due to the complex disturbance from other sources. This paper proposes a heterogeneous deep model fusion (HDMF) method to solve the problem in a unified framework. The contributions include: 1) The convolutional neural network (CNN) and long short-term memory (LSTM) are combined by two different ways without prior knowledge involved; 2) A large database, including eleven types of single-carrier modulation signals with various noises as well as a fading channel, is collected with various signal-to-noise ratios (SNRs) based on a real geographical environment; and 3) Experimental results demonstrate that HDMF is super capable of copping with the AMC problem, and achieves much better performance when compared with the independent network. The source code and the database will be publically available.

The automatic diagnosis of cardiovascular diseases has received much attention in the deep learning field. In this context, the segmentation of the left ventricle endocardium constitutes a major task in diagnosing heart conditions such as health failure and hypertrophic cardiomyopathy. The objective of this paper is to propose a "deep convolution network" for segmenting the internal cavity of the left ventricle (endocardium) using MRI images. In particular, we design an improved UNET model which handles additional inception modules for efficiently segmenting the internal cavity of the left ventricle. Our approach has been validated on the Sunnybrook Cardiac Data (SCD) dataset and has showed promising results in terms of precision. More specifically, the improved UNET largely outperforms the baseline UNET model and many existing state-of-the-art methods.

Fault detection is an important step for subsurface interpretation and reservoir characterization from 3D seismic images. Due to the numerous and complicated faulting structures of seismic images, manual seismic interpretation is time taking and need intensive work. To overcome this problem, geoscientists are coming up with productive computer-aided techniques for assisting in interpreter science for many years. However, in this paper, we used a pre-trained CNN model to predict faults from the 3D seismic volume of the Poseidon field of Browse Basin, Australia. Basically, this field is highly structured with complex normal faulting throughout the targeted Plover Formations. So, our motivation for this work in this field is to compare machine learning-based fault prediction to user interpreted faults identification with supported by seismic variance attribute. We found very satisfying result using DL with having an improved fault probability volume that outperform variance technology. Therefore, we propose that this workflow could reduce time and able to predict fault quite accurately in any field area.

When extract building from high resolution remote sensing image with meter/sub-meter accuracy, the shade of trees and interference of roads are the main factors of reducing the extraction accuracy. Proposed a Bayesian Convolutional Neural Networks(BCNET) model base on standard fully convolutional networks(FCN) to solve these problems. First take building with no shade or artificial removal of shade as Sample-A, woodland as Sample-B, road as Sample-C. Set up 3 sample libraries. Learn these sample libraries respectively, get their own set of feature vector; Mixture Gauss model these feature vector set, evaluate the conditional probability density function of mixture of noise object and roofs; Improve the standard FCN from the 2 aspect:(1) Introduce atrous convolution. (2) Take conditional probability density function as the activation function of the last convolution. Carry out experiment using unmanned aerial vehicle(UVA) image, the results show that BCNET model can effectively eliminate the influence of trees and roads, the building extraction accuracy can reach 97%.

The continuous and precise mapping of glacier calving fronts is essential for monitoring and understanding rapid glacier changes in Antarctica and Greenland, which have the potential for significant sea level rise within the current century. This effort has been mostly restricted to the slow and painstaking manual digitalization of the calving front positions in thousands of satellite imagery products. Here, we have developed a machine learning toolkit to robustly and automatically detect glacier calving front margins in satellite imagery. The toolkit is based on semantic image segmentation using Convolutional Neural Networks (CNN) with a modified U-Net architecture to isolate the calving fronts from satellite images after having been trained with a dataset of images and their corresponding manually-determined calving fronts. As a case study we train our neural network on a varied set Landsat images with lowered resolutions from Jakobshavn, Sverdrup, and Kangerlussuaq glaciers, Greenland and test the results on novel images from Helheim glacier, Greenland to evaluate the performance of the approach. The neural network is able to identify the calving front in new images with a mean deviation of 96.3 m from the true fronts, equivalent to 1.97 pixels on average, while the corresponding error for manually-determined fronts on the same resolution images is 92.5 m. We find that the trained neural network significantly outperforms common edge detection techniques, and can be used to continuously map out calving-ice fronts with a variety of data products.

Alzheimer's disease (AD) is a progressive and evolving neurodegenerative disease with an insidious onset that can lead to memory loss and cognitive impairment. There is no effective treatment for this disease. However, early diagnosis plays an important role in treatment planning to slow down its progression, as treatment has the greatest impact in the early stages of the disease. Neurological images obtained through different imaging techniques provide powerful information and help diagnose the disease. With the wide application of deep learning techniques in disease diagnosis, especially the prominence of Convolutional Neural Networks (CNNs) in computer vision and image processing, more and more studies are proposing the use of this algorithm for the diagnosis of AD. In this paper, we first systematically introduce the impact of AD on people, detailing the biomarkers, early clinical symptoms, and risk factors of this disease. Secondly, it goes on to detail the development of CNNs, their form, and methods to help diagnose AD. It is proposed that CNNs can help diagnose AD by analyzing medical imaging data, particularly structural brain scans such as magnetic resonance imaging (MRI) and functional scans such as positron emission tomography (PET). Finally, it is concluded that CNNs are of great importance for the diagnosis of AD and that they are likely to play an increasingly important role in the early detection of the disease, the understanding of disease mechanisms, and ultimately, in the development of effective AD therapies and interventions. CNNs are playing an increasingly important role in the Their potential impact on healthcare emphasizes the importance of continued research and innovation in neural networks and medical imaging.

With the development of deep learning technology, various structures and research methods for super-resolution restoration of natural images and document images have been introduced. In particular, a number of recent studies have been conducted and developed in image restoration using generative adversarial network. Super-resolution restoration is ill-posed problem because of some complex restraints such as a lot of high-resolution images being restored for the same low-resolution image and also difficulty in restoring noises like edges, light smudging, and blurring. In this study, we utilized the spatially adaptive de-normalization (SPADE) structure for document image restoration to solve previous problems such as edge unclearness, hardness to catch features of texts, and the image color transition. Consequently, it can be confirmed that the edge of the character and the ambiguous stroke are restored more clearly when contrasting with the other previously suggested methods. Also, the proposed method’s PSNR and SSIM scores are geting 8% and 15% higher, respectively, compared to the previous methods.

Tail-biting convolutional codes extend the classical zero-termination convolutional codes: Both encoding schemes force the equality of start and end states, but under the tail-biting each state is a valid termination. This paper proposes a machine-learning approach to improve the state-of-the-art decoding of tail-biting codes, focusing on the widely employed short length regime as in the LTE standard. This standard also includes a CRC code. First, we parameterize the circular Viterbi algorithm, a baseline decoder that exploits the circular nature of the underlying trellis. An ensemble combines multiple such weighted decoders, each decoder specializes in decoding words from a specific region of the channel words' distribution. A region corresponds to a subset of termination states; the ensemble covers the entire states space. A non-learnable gating satisfies two goals: it filters easily decoded words and mitigates the overhead of executing multiple weighted decoders. The CRC criterion is employed to choose only a subset of experts for decoding purpose. Our method achieves FER improvement of up to 0.75dB over the CVA in the waterfall region for multiple code lengths, adding negligible computational complexity compared to the circular Viterbi algorithm in high SNRs.

In the field of maneuvering target tracking, the combined observations of azimuth and Doppler may cause weak observation or non-observation in the application of traditional target tracking algorithms. Additionally, traditional target-tracking algorithms require pre-defined multiple mathematical models to accurately capture the complex motion states of targets, while model mismatch and unavoidable measurement noise lead to significant errors in target state prediction. To address those above challenges, in recent years, the target-tracking algorithms based on neural networks, such as recurrent neural networks (RNNs), long short-term memory (LSTM) networks, and Transformer architectures, have been widely used for their unique advantages to achieve accurate predictions. To better model the nonlinear relationship between the observation time series and the target state time series, as well as the contextual relationship among time series points, we present a deep learning algorithm called recursive downsample-convolve-interact neural network (RDCINN) based on convolutional neural network (CNN) that downsamples time series into sub-sequences and extracts multi-resolution features to enable the modeling of complex relationships between time series, which overcomes the shortcomings of traditional target-tracking algorithms in using observation information inefficiently due to weak observation or non-observation. The experimental results show that our algorithm outperforms other existing algorithms in the scenario of strong maneuvering target tracking with the combined observations of azimuth and Doppler.

In the age of social media, images come from unreliable sources that may change their content to fit a narrative. Automated detection of forged images is a complex task, especially considering recent technological advances in image manipulation software. However, in the existing literature, it is possible to identify two general approaches adopted by detection methods, active and passive. Active techniques preemptively act on an image by inserting structures before any manipulation is made that can verify its authenticity. In contrast, passive methods analyze the content of an image in search of traces of manipulation. In this way, this research proposes a novel approach to image manipulation detection, combining two passive methods through neural networks, creating a generalist approach capable of detecting with greater accuracy than the methods that compose it. Furthermore, this work used a combination of four datasets available in the literature for training and evaluation. After training, the merged approach obtained an accuracy of 89.59% in the set of validation images, significantly higher than the model trained with only unaltered images, which obtained 78.64%, and the two other models trained using images with a feature selection method applied to enhance inconsistencies that obtained 68.02% for Error-Level Analysis images and 50.70% for the method using Discrete Wavelet Transform. In addition to the performance improvement, the proposed approach’s accuracy variation was lower than the other models.

Hospital readmission and length of stay prediction provide info to manage hospitals’ bed capacity and the number of required staff, especially during pandemics. We present a hybrid deep model called Genetic Algorithm-Optimized Convolutional Neural Network (GAOCNN) with a unique preprocessing method to predict hospital readmission and the length of stay in patients having various conditions. GAOCNN uses one-dimensional convolutional layers to predict hospital readmission and length of stay. The parameters of the layers are optimized using a genetic algorithm. To show the performance of the proposed model in patients with various conditions, we evaluate the model under three healthcare datasets; the Diabetes 130-US hospitals dataset, the COVID-19 dataset, and the MIMIC-III dataset. The diabetes 130-US hospitals dataset has information on both readmission and the length of stay, while COVID-19 and MIMIC-III datasets just include information on the length of stay. Experimental results show that the proposed model’s accuracy for hospital readmission is 97.2% for diabetic patients. Also, the accuracy of the length of stay prediction is 89%, 99.4%, and 94.1% for diabetic, COVID-19, and ICU patients, respectively. These results confirm the superiority of the proposed model compared to existing methods. Our findings offer a platform for managing healthcare funds and resources for patients with various diseases.

The small satellite market continues to grow year after year. A compound annual growth rate of 17% is estimated during the period between 2020 and 2025. Low-cost satellites can send a vast amount of images to be post-processed at the ground to improve the quality and extract detailed information. In this domain lies the resolution enhancement task, where a low-resolution image is converted to a higher resolution automatically. Deep learning approaches to Super-Resolution (SR) reached the state-of-the-art in multiple benchmarks; however, most of them were studied in a single-frame fashion. With satellite imagery, multi-frame images can be obtained at different conditions giving the possibility to add more information per image and improve the final analysis. In this context, we developed and applied to the PROBA-V dataset of multi-frame satellite images a model that recently topped the European Space Agency’s Multi-frame Super Resolution (MFSR) competition. The model is based on proven methods that worked on 2D images tweaked to work on 3D: the Wide Activation Super Resolution (WDSR) family. We show that with a simple 3D CNN residual architecture with WDSR blocks and a frame permutation technique as data augmentation better scores can be achieved than with more complex models. Moreover, the model requires few hardware resources, both for training and evaluation, so it can be applied directly from a personal laptop.

Human action recognition (HAR) based on skeleton data is a challenging yet important technique because of its wide range of applications in many fields, including patient monitoring, security surveillance, and observing human-machine interactions. Many algorithms that attempt to distinguish between many types of activities have been proposed. However, most practical applications require highly accurate detection of specific types of activities. In this study, a novel and highly accurate spatiotemporal graph autoencoder network for HAR based on skeleton data is proposed. Furthermore, an extensive study was conducted using different modalities. For this purpose, a spatiotemporal graph autoencoder that automatically learns spatial as well as temporal patterns from human skeleton datasets was built. The powerful graph convolutional network named GA-GCN, developed in this study, notably outperforms most of the existing state-of-the-art methods based on two common datasets, namely NTU RGB+D and NTU RGB+D 120 . On the first dataset, we achieved an accuracy of 92.3% and 96.7% based on the cross-subject and cross-view evaluations, respectively. On the other more challenging dataset (i.e. NTU RGB+D 120), GA-GCN achieved 88.8% and 90.4% based on the cross-subject and cross-set evaluation, respectively.

Speech Emotion Recognition (SER) is an interesting and difficult problem to handle. In this paper, we deal with it through the implementation of deep learning networks. We have designed and implemented six different deep learning networks, a Deep Belief Network (DBN), a simple deep neural network (SDNN), a LSTM network (LSTM), a LSTM network with the addition of an attention mechanism (LSTM-ATN), a Convolutional neural network (CNN), and a Convolutional neural network with the addition of an attention mechanism (CNN-ATN), having in mind, apart from solving the SER problem, to test the impact of attention mechanism to the results. Dropout and Batch Normalization techniques are also used to improve the generalization ability (prevention of overfitting) of the models as well as to speed up the training process. The Surrey Audio-Visual Expressed Emotion database (SAVEE), and the Ryerson Audio-Visual Database (RAVDESS) database were used for training and evaluation of our models. The results showed that networks with the addition of the attention mechanism did better than the others. Furthermore, they showed that CNN-ATN was the best among tested networks, achieving an accuracy of 74% for the SAVEE and 77% for the RAVDESS dataset, and exceeded existing state-of-the-art systems for the same datasets.

Due to the recent trend of unmanned economy, retail stores have gradually reduced their service and cashier manpower. The retail product recognition becomes one of key issues for unmanned shopping. Although the success of deep neural networks has made object recognition feasible in a variety of applications, it still struggles to perform well on a large number of classes of retail products. In this paper, we propose an improved ensemble learning-based approach for retail product recognition. In the proposed approach, the object classification network is first improved through feature extraction and block attention. An ensemble model is then built by integrating multiple network models and using loss selection as model weights. In the experiments, the feasibility of our ensemble learning-based approach has been evaluated through many production items. The results demonstrate the effectiveness of the proposed approach compared with previous retail product recognition methods.

The Sentinel-2 satellite mission offers high resolution multispectral time series image data, enabling the production of detailed land cover maps globally. At this scale, the trade-off between processing time and result quality is a central design decision. Currently, this machine learning task is usually performed using pixelwise classification methods. The radical shift of the computer vision field away from hand engineered image features and towards more automation by representation learning comes with many promises, including higher quality results and less engineering effort. In this paper we assess fully convolutional neural networks architectures as replacements for a Random Forest classifier in an operational context for the production of high resolution land cover maps with Sentinel-2 time series at the country scale. Our contributions include a framework for working with Sentinel-2 L2A time series image data, an adaptation of the U-Net model for dealing with sparse annotation data while maintaining high resolution output, and an analysis of those results in the context of operational production of land cover maps.

Convolutional Neural Networks combined with autonomous drones are increasingly seen as enablers of partially automating the aircraft maintenance visual inspection process. Such an innovative concept can have a significant impact on aircraft operations. Through supporting aircraft maintenance engineers detect and classify a wide range of defects, the time spent on inspection can significantly be reduced. Examples of defects that can be automatically detected include aircraft dents, paint defects, cracks and holes, and lightning strike damage. Additionally, this concept could also increase the accuracy of damage detection and reduce the number of aircraft inspection incidents related to human factors like fatigue and time pressure. In our previous work, we have applied a recent Convolutional Neural Network architecture known by MASK R-CNN to detect aircraft dents. MASK-RCNN was chosen because it enables the detection of multiple objects in an image while simultaneously generating a segmentation mask for each instance. The previously obtained F1 and F2 scores were 62.67% and 59.35% respectively. This paper extends the previous work by applying different techniques to improve and evaluate prediction performance experimentally. The approaches uses include (1) Balancing the original dataset by adding images without dents; (2) Increasing data homogeneity by focusing on wing images only; (3) Exploring the potential of three augmentation techniques in improving model performance namely flipping, rotating, and blurring; and (4) using a pre-classifier in combination with MASK R-CNN. The results show that a hybrid approache combining MASK R-CNN and augmentation techniques leads to an improved performance with an F1 score of (67.50%) and F2 score of (66.37%)

Wet granulation is a frequent process in the pharmaceutical industry. As a starting point for numerous dosage forms, the quality of the granulation not only affects subsequent production steps but also impacts the quality of the final product. It is thus crucial and economical to monitor this operation thoroughly. Here, we report on identifying different phases of a granulation process using a machine learning approach. The phases reflect the water content which in turn influences the processability and quality of the granule mass. We used two kinds of microphones and an acceleration sensor to capture acoustic emissions and vibrations. We trained convolutional neural networks (CNNs) to classify the different phases using transformed sound recordings as the input. We achieved a classification accuracy of up to 90 % using vibrational data and an accuracy of up to 97 % using the audible microphone data. Our results indicate the suitability of using audible sound and machine learning to monitor pharmaceutical processes. Moreover, since recording acoustic emissions is contactless, it readily complies with legal regulations and present Good Manufacturing Practices.

As a relatively high percentage of adenoma polyps are missed, a computer-aided diagnosis (CAD) tool based on deep learning can aid the endoscopist in diagnosing colorectal polyps or colorectal cancer in order to decrease polyps missing rate and prevent colorectal cancer mortality. Convolutional Neural Network (CNN) is a deep learning method and has achieved better results in detecting and segmenting specific objects in images in the last decade than conventional models such as regression, support vector machines or artificial neural networks. In recent years, based on the studies in medical imaging criteria, CNN models have acquired promising results in detecting masses and lesions in various body organs, including colorectal polyps. In this review, the structure and architecture of CNN models and how colonoscopy images are processed as input and converted to the output are explained in detail. In most primary studies conducted in the colorectal polyp detection and classification field, the CNN model has been regarded as a black box since the calculations performed at different layers in the model training process have not been clarified precisely. Furthermore, I discuss the differences between the CNN and conventional models, inspect how to train the CNN model for diagnosing colorectal polyps or cancer, and evaluate model performance after the training process.

Although automated Acute Lymphoblastic Leukemia (ALL) detection is essential, it is challenging due to the morphological correlation between malignant and normal cells. The traditional ALL classification strategy is arduous, time-consuming, often suffers inter-observer variations, and necessitates experienced pathologists. This article has automated the ALL detection task, employing deep Convolutional Neural Networks (CNNs). We explore the weighted ensemble of deep CNNs to recommend a better ALL cell classifier. The weights are estimated from ensemble candidates' corresponding metrics, such as accuracy, F1-score, AUC, and kappa values. Various data augmentations and pre-processing are incorporated for achieving a better generalization of the network. We train and evaluate the proposed model utilizing the publicly available C-NMC-2019 ALL dataset. Our proposed weighted ensemble model has outputted a weighted F1-score of 88.6%, a balanced accuracy of 86.2%, and an AUC of 0.941 in the preliminary test set. The qualitative results displaying the gradient class activation maps confirm that the introduced model has a concentrated learned region. In contrast, the ensemble candidate models, such as Xception, VGG-16, DenseNet-121, MobileNet, and InceptionResNet-V2, separately produce coarse and scatter learned areas for most example cases. Since the proposed ensemble yields a better result for the aimed task, it can experiment in other domains of medical diagnostic applications.

Automated lip reading is a research problem that has developed considerably in recent years. Lip reading is evaluated both visually and audibly in some cases. The lip reading model is a field of use for detecting specific words using images from security cameras, but it is not possible to use audio-visual databases in this situation. It is not possible to obtain the sound input of the pronounced word in all cases. We collected a new Turkish dataset with only the image in this study. The new dataset is produced using Youtube videos, which is an uncontrolled environment. For this reason, images have difficult parameters in terms of environmental factors such as light, angle, color, and personal characteristics of the face. Despite the different features on the human face such as mustache, beard, and make-up, the visual speech recognition problem was developed on 10 classes including single words and two-word phrases using Convolutional Neural Networks (CNN) without any intervention on the data. The proposed study using only-visual data obtained a model which is automated visual speech recognition with a deep learning approach. In addition, since this study uses only-visual data, the computational cost and resource usage is less than in multi-modal studies. It is also the first known study to address the lip reading problem with a deep learning algorithm using a new dataset belonging to the Ural-Altaic languages.

This paper presents a non-random weight initialization method in convolutional layers of neural networks examined with the Fast Gradient Sign Method (FSGM) attack. This paper's focus is convolutional layers, and are the layers that have been responsible for better than human performance in image categorization. The proposed method induces earlier learning through the use of striped forms, and as such has less unlearning of the existing random number speckled methods, consistent with the intuitions of Hubel and Wiesel. The proposed method provides a higher performing accuracy in a single epoch, with improvements of between 3-5% in a well known benchmark model, of which the first epoch is the most relevant as it is the epoch after initialization. The proposed method is also repeatable and deterministic, as a desirable quality for safety critical applications in image classification within sensors. That method is robust to Glorot/Xavier and He initialization limits as well. The proposed non-random initialization was examined under adversarial perturbation attack through the FGSM approach with transferred learning, as a technique to measure the affect in transferred learning with controlled distortions, and finds that the proposed method is less compromised to the original validation dataset, with higher distorted datasets.

The semantic segmentation of high-resolution remote sensing images (HR-RSI) is crucial for a wide range of applications, such as precision agriculture, urban planning, natural resource assessment, and ecological monitoring. However, accurately classifying pixels in HR-RSI faces challenges due to densely distributed small objects and scale variations. Existing techniques, including Convolutional Neural Networks (CNNs) and other methods for hierarchical feature extraction and fusion of remote sensing image, often do not achieve the desired accuracy. In this paper, we propose a novel approach called the Hierarchical Rich-scale Fusion Network (HRFNet) to address these challenges. HRFNet utilizes advanced information rating and image partition techniques to extract rich-scale features within image layers. This allows for the adaptive exploration of both local and global contextual information. Moreover, we introduce a structured intra-layer to inter-layer feature aggregation module, which enables the adaptive extraction of fine-grained details and high-level semantic information from multi-layer feature maps in a highly flexible manner. Extensive experimentation has been conducted to validate the effectiveness of our proposed method. Our results demonstrate that HRFNet outperforms existing techniques, achieving state-of-the-art (SOTA) results on benchmark datasets, specifically the ISPRS Potsdam and Vaihingen datasets.

Artificial Intelligence (AI) made its first appearance in 1956, and since then it has progressively introduced itself in healthcare system and patients’ information and care. AI functions can be grouped under the following headings: Machine Learning (ML), Deep Learning (DL), Artificial Neural Network (ANN), Convolutional Neural Network (CNN), Computer Vision (CV). Chronic limb-threatening ischemia (CLTI) represents the last stage of peripheral artery disease (PAD), and has increased over recent years, together with the rise in prevalence of diabetes and population ageing. Nowadays AI grants the possibility of developing new diagnostic and treatment solutions in the vascular field, given the possibility of accessing clinical, biological, and imaging data. By assessing the vascular anatomy in every patient, as well as the burden of atherosclerosis, and classifying the level and degree of disease, sizing and planning the best endovascular treatment, defining the perioperative complications risk, integrating experiences and resources between different specialties, identifying latent PAD, thus offering evidence-based solutions and guiding surgeons in the choice of the best surgical technique, AI challenges the role of the physician’s experience in PAD treatment.

Even today, the automatic digitisation of scanned documents in general but especially the automatic optical music recognition (OMR) of historical manuscripts still remain an enormous challenge, since both handwritten musical symbols and text have to be identified. This paper focuses on the Medieval so-called square notation developed in the 11th-12th century, which is already composed of staff lines, staves, clefs, accidentals, and neumes, that are roughly spoken connected single notes. The aim is to develop an algorithm that captures both the neume and pitch, that is melody information that can be used to reconstruct the original writing. Our pipeline is similar to the standard OMR approach and comprises a novel staff line and symbol detection algorithm, based on deep Fully Convolutional Networks (FCN), which perform pixel-based predictions for either staff lines or symbols and their respective types. Then, the staff line detection combines the extracted lines to staves and yields an F1-score of over 99% for both detecting lines and complete staves. For the music symbol detection we choose a novel approach that skips the step to identify neumes and instead directly predicts note components (NCs) and their respective affiliation to a neume. Furthermore, the algorithm detects clefs and accidentals. Our algorithm recognises these symbols with an F1-score of over 96% if the type is ignored and predicts the true symbol sequence of a staff with a diplomatic symbol accuracy rate (dSAR) of about 87%. If only the NCs without their respective connection to a neume, all clefs, and accidentals are of interest the algorithm reaches an harmonic symbol accuracy rate (hSAR) of approximately 90%.

Land Surface Temperature (LST) is significant for climatological and environmental studies. LST products from satellites, however, suffer from the tradeoff between spatial and temporal resolution. Spatial downscaling has emerged as a well explored field aiming to overcome limitations arising from this tradeoff. Previous research on regression based LST downscaling models focused on utilizing predictors derived from optical imagery. Weather-dependency of optical imagery data, however, can influence downscaling models by the weather conditions. To cope this issue, in this study, we involve predictors derived from the weather-independent Sentinel-1 Synthetic Aperture Radar (SAR) imagery to downscale Landsat-8 LST data. In this context, we propose to use machine learning techniques, namely Random Forest (RF) and Convolutional Neural Networks (CNN). To demonstrate the applicability and performance of the proposed method, extensive experimental analyses were conducted over Zuid-Holland in the Netherlands. From the experiments, we found that the results obtained with radar predictors were comparable to those achieved using optical predictors. This confirms that the proposed method indeed paves a new way for mapping land surface temperature using SAR images.

Radio Frequency Interference (RFI) detection and characterization play a critical role to in ensuring the security of all wireless communication networks. Advances in Machine Learning (ML) have led to the deployment of many robust techniques dealing with various types of RFI. To sidestep an unavoidable complicated feature extraction step in ML, this paper proposes an efficient end-to-end method using the latest advances in deep learning to extract the appropriate features of the RFI signal. Moreover, this study utilizes the benefits of transfer learning to determine both the type of received RFI signals and their modulation types. To this end, the scalogram of the received signals is used as the input of the pre-trained convolutional neural networks (CNN), followed by a fully-connected classifier. This study considers a digital video stream as the signal of interest (SoI), transmitted in a real-time satellite-to-ground communication using DVB-S2 standards. To create the RFI dataset, the SoI is combined with three well-known jammers namely, continuous-wave interference (CWI), multi- continuous-wave interference (MCWI), and chirp interference (CI). This study investigated four well-known pre-trained CNN architectures, namely, AlexNet, VGG-16, GoogleNet, and ResNet-18, for the feature extraction to recognize the visual RFI patterns directly from pixel images with minimal preprocessing. Moreover, the robustness of the proposed classifiers is evaluated by the data generated at different signal to noise ratios (SNR).

Detailed flow distributions in vascular systems are the key to identifying hemodynamic risk factors for the development and progression of vascular diseases. Although computational fluid dynamics (CFD) has been widely used in bioengineering research on hemodynamics predictions, not only are high-fidelity CFD simulations time-consuming and computing-expensive, but also not friendly to clinical applications due to the difficulty of comprehensive numerical calculations. Machine learning (ML) algorithums to estimate the flow field in vascular systems based on the angiographic images of the blood flow using existed diagnostic tools are emerging as a new pathway to facilitate the mapping of hemodynamics. In present work, the dye injection in a water flow was simulated as an analogy of the contrast perfusion in blood flow using CFD. In the simulation, the light passes through the flow field and generates projective images, as an analogy of X-ray imaging. The simulations provide both the ground truth velocity field and the projective images of the flow with dye patterns. A rough velocity field was estimated using the optical flow method (OFM) based on projective images. ML algorithums are then trained using the ground truth CFD data and the OFM velocity estimation as the input. Finally, the interpretable (logistic regression) and deep (neural networks, convolutional neural networks, long short term memory) machine learning models are validated by using parallel in vitro experiments on the same flow setup. The validation results showed that the employed ML model significantly reduced the error rate from 53.5% to 2.5% in average for the v-velocity estimation.

Accurate and efficient medicinal plant image classification is of utmost importance as these plants produce a wide variety of bioactive compounds that offer therapeutic benefits. With a long history of medicinal plant usage, different parts of plants, such as flowers, leaves, and roots, have been recognized for their medicinal properties and are used for plant identification. However, leaf images are extensively used due to their convenient accessibility and are a major source of information. In recent years, transfer learning and fine-tuning, which use pre-trained deep convolutional networks to extract pertinent features, has emerged as an extremely effective approach for image identification problems. This study is leveraging the power of three component deep convolutional neural networks, namely VGG16, VGG19 and DenseNet201, to derive features from the input images of the medicinal plant dataset, containing leaf images of 30 classes. The models were compared and ensembled to make four hybrid models to enhance the predictive performance by utilizing the averaging and weighted averaging strategies. Quantitative experiments were carried out to evaluate the models on the Mendeley medicinal leaf dataset. The resultant ensemble of VGG19+DensNet201 with fine-tuning showcases enhanced capability in identifying medicinal plant images with an improvement of 7.43% and 5.8% compared with VGG19 and VGG16. Furthermore, VGG19+DensNet201 can outperform its standalone counterparts by achieving an accuracy of 99.12% on the test set. A thorough assessment with metrics such as accuracy, recall, precision, and f1-score firmly establishes the effectiveness of the ensemble strategy.

The interpretation of land use and land cover (LULC) is an important issue in the fields of high-resolution remote sensing (RS) image processing and land resource management. Fully training a new or existing convolutional neural network (CNN) architecture for LULC classification requires a large amount of remote sensing images. Thus, fine-tuning a pre-trained CNN for LULC detection is required. To improve the classification accuracy for high resolution remote sensing images, it is necessary to use another feature descriptor and to adopt a classifier for post-processing. A fully connected conditional random fields (FC-CRF), to use the fine-tuned CNN layers, spectral features, and fully connected pairwise potentials, is proposed for image classification of high-resolution remote sensing images. First, an existing CNN model is adopted, and the parameters of CNN are fine-tuned by training datasets. Then, the probabilities of image pixels belong to each class type are calculated. Second, we consider the spectral features and digital surface model (DSM) and combined with a support vector machine (SVM) classifier, the probabilities belong to each LULC class type are determined. Combined with the probabilities achieved by the fine-tuned CNN, new feature descriptors are built. Finally, FC-CRF are introduced to produce the classification results, whereas the unary potentials are achieved by the new feature descriptors and SVM classifier, and the pairwise potentials are achieved by the three-band RS imagery and DSM. Experimental results show that the proposed classification scheme achieves good performance when the total accuracy is about 85%.

Iron ore processing involves critical steps that affect the quality of the final product. Determining the diameter of the pellets is the necessary initial step for volume measurement and, ultimately, for porosity and bulk density measurement, crucial characteristics for optimizing the burning process in the blast furnace. Traditional measurement methods using mercury present issues related to operator safety and environmental preservation, while the practices described in ISO 4698 standard require lengthy preparation and execution time. In light of environmental needs, operator safety, and the time consumed in tests, implementing a new method based on digital image processing and convolutional neural networks for measuring the diameter of burnt iron ore pellets is proposed. The methodology involves capturing images of the pellets using a high-resolution camera and utilizing digital image processing and neural networks capable of performing pixel-by-pixel object segmentation in the images, providing precise information about the pellets to calculate their volume automatically. The results were compared with those obtained using traditional methods, evaluating their conformity with the ISO 4698 standard. In conclusion, this study offers a new approach to measuring the volume of burnt iron ore pellets, providing accurate and reliable results in compliance with safety and environmental preservation standards.