In high-traffic harbor waters, marine radar frequently encounters signal interference stemming from various obstructive elements, thereby presenting formidable obstacles in the precise identification of ships. To achieve precise pixel-level ship identification in the complex environments, a customized neural network-based ship segmentation algorithm named MrisNet is proposed. MrisNet employs a lightweight and efficient FasterYOLO network to extract features from radar images at different levels, capturing fine-grained edge information and deep semantic features of ship pixels. To address the limitation of deep features in the backbone network lacking detailed shape and structured information, an adaptive attention mechanism is introduced after the FasterYOLO network to enhance crucial ship features. To fully utilize the multi-dimensional feature outputs, MrisNet incorporates a Transformer structure to reconstruct the PANet feature fusion network, allowing for the fusion of contextual information and capturing more essential ship information and semantic correlations. In the prediction stage, MrisNet optimizes the target position loss using the EIoU function, enabling the algorithm to adapt to ship position deviations and size variations, thereby improving segmentation accuracy and convergence speed. Experimental results demonstrate that MrisNet achieves high recall and precision rates of 94.8% and 95.2%, respectively, in ship instance segmentation, outperforming various YOLO and other single-stage algorithms. Moreover, MrisNet has a model parameter size of 13.8M and real-time computational cost of 23.5G, demonstrating notable advantages in terms of convolutional efficiency. In conclusion, MrisNet accurately segments ships with different spot features and under diverse environmental conditions in marine radar images. It exhibits outstanding performance, particularly in extreme scenarios and challenging interference conditions, showcasing robustness and applicability.

Mindful Eating (ME) is an increasingly popular dietary practice that may mediate the sympto-matology of Eating Disorders (EDs). Orthorexia Nervosa (ON) is a recently proposed eating dis-order characterized by excessive concern for healthy eating and calorie control. The purpose of this study is to investigate the relationship between ME and Psychological Distress (PD) with EDs and ON. In a cross-sectional study involving 407 adults, this research explored the interplay between ME, PD, and EDs, notably ON. Utilizing validated scales for ME (MES-16), EDs (EDE-QS), ON (ORTO-R), and PD, subcategorized in depression, anxiety, and stress (DASS-21), alongside inquiries into Instagram usage, significant correlations emerged. ME demonstrated a robust negative asso-ciation with both EDE-QS (r=-0.501, p<0.001) and ORTO-R (r=-0.519, p<0.001), while EDE-Q and ORTO-R showed significant correlations with depression, anxiety, and stress (p<0.001). Variations in ORTO-R mean scores were notably distinct across different ME categories (p<0.001), emphasizing the relationship between ME and ON. Additionally, participants with extensive daily Instagram use displayed significantly higher ON scores compared to minimal users (p<0.05). These findings sug-gest a negative link between ME with EDs and ON, highlighting ME's potential role in reducing ON tendencies. Furthermore, Instagram use seems associated with PD and elevated ON levels.

This proposed research explores a novel approach to image classification by deploying a complex-valued neural network (CVNN) on a field-programmable gate array (FPGA), specifically for classifying 2D images transformed into polar form. The aim of this research is to address the limitations of existing neural network models in terms of energy and resource efficiency, by exploring the potential of FPGA-based hardware acceleration in conjunction with advanced neural network architectures like CVNNs. The methodological innovation of this research lies in the Cartesian to polar transformation of 2D images, effectively reducing the input data volume required for neural network processing. Subsequent efforts focused on constructing a CVNN model optimized for FPGA implementation, emphasizing the enhancement of computational efficiency and overall performance. The experimental findings provide empirical evidence supporting the efficacy of the image classification system developed in this study. One of the developed models, CVNN_128, achieves an accuracy of 88.3% with an inference time of just 1.6ms and a power consumption of 4.66mW for the classification of the MNIST test dataset consists of 10,000 frames. While there is a slight concession in accuracy compared to recent FPGA implementations that achieve 94.43%, our model significantly excels in classification speed and power efficiency—surpassing existing models by more than a factor of 100. In conclusion, the paper demonstrates the substantial advantages of FPGA-implementation of CVNNs for image classification tasks, particularly in scenarios where speed, resource, and power consumption are critical. The study’s reproducible results and corresponding code are available on GitHub at the following link: https://github.com/mahmad2005/CVNNonFPGA

Malaria is one of the life-threatening disease caused by the Anopheles mosquitoes affecting the human red blood cells. It is the global health concern which is quite common in tropical as well as sub-tropical parts of the World. Therefore, it is an important to have an effective computer aided system in place for early detection and treatment. State of the art machine learning techniques are used to detect the infected malaria cells from the pool of images. As the visual heterogeneity of the malaria dataset is highly complex and dynamic, therefore higher number of images are needed to train the machine learning models effectively. However, hospitals as well as medical institutions do not share the medical image data for collaboration due to general data protection regulation (GDPR) and data protection act (DPA). To overcome this collaborative challenge, our research is inspired to use the real-time medical image data while using the framework of federated learning (FL) framework that will ensure the data privacy while mutual collaboration of hospitals and medical institutes. We have used the state of the art machine learning models that include the Resnet50 and densenet in a federated learning framework. We have experimented both models in different settings and our preliminary results showed that the densenet model performed better in accuracy (75%) in contrast to resnet50 (72%) while considering 8 clients, while the trend is observed common in 4 clients with the similar accuracy of 94% and 6 client showed that the densenet model performed quite well with the accuracy of 92% while resnet50 achieving only 72%.

Techniques and tools meant to aid fire management activities in the Cerrado, such as accurately determining fuel load and composition spatially and temporally, are pretty scarce. The need to have fuel information for more efficient management in a considerably heterogeneous, biodiverse, and fire-dependent environment makes a constant search for the improvement of the use of remote sensing techniques in determining fuel characteristics. This study presents the following objectives: (1) to assess the use of data from Landsat 8 OLI images to estimate the fine surface fuel load of the Cerrado during the dry season by adjusting multiple linear regression equations; (2) to estimate fuel load through Random Forest and k-Nearest Neighbor (k-NN) algorithms in comparison with re-gression analyses; and (3) to evaluate the importance of predictor variables from satellite images. Therefore, 64 sampling units were collected, and the pixel values associated with the field plots were extracted in a 3 x 3 pixel window surrounding the reference pixel. For the multiple linear regression analyses, the R² values ranged from 0.63 to 0.78, and the models fitted by the Random Forest al-gorithm ranged from 0.52 to 0.83, while those fitted by k-NN ranged from 0.30 to 0.68. Adopting the Random Forest algorithm resulted in improvements in the statistical metrics of evaluation of the fuel load estimates for Cerrado grassland relative to the multiple linear regression analyses.

Abstract: The two-dimensional Fuzzy Transform was applied in image compression. The quality of the image compressed with this methodis better than that obtained using methods based on fuzzy relation equations and comparable with that of the JPEG method, with better compression execution times. In this paper we propose a variation of the method based on the two-dimensional Fuzzy Transform of image compression, in which the image is first partitioned into blocks and then each block is compressed with the higher compression rate. The advantage of this method consistsinto a greater compression of the image, guaranteeing a high quality of the reconstructed image.The results show that our method is better than the Fuzzy Transform method improving also the quality of the reconstructed image.

The utilization of tempered blast furnace (BF) slag through the direct fiber forming process to create high-value thermal insulation materials offers a dual benefit: it efficiently harnesses the latent heat within unused slag and substantially enhances the value of blast-furnace slag utilization. However, gauging the melting properties of iron slag under high temperatures is a challenge. In this study, we explore the melting behavior of SiO2 within a high-temperature molten pool. We employ dynamic visual data (video stream) captured via a non-contact charge coupled device (CCD) video recording system to extract SiO2 contours through image processing. The change in image centroid characteristics is used to establish a convolution function relationship, and MATLAB's traversal search algorithm determines SiO2's centroid position. Given that SiO2 is proportionate to crucible pixels, the area of SiO2 is calculated through pixel statistics within these contours. Subsequently, we propose a new indirect method to process image information, yielding SiO2 volume and mass at different time points. An exponential fitting yields the melting rate function of SiO2. Finally, we compare this indirect method with shape from shading (SFS), quantitative characterization, and dimensional analysis techniques. We also discuss the strengths and limitations of each method. Our findings reveal that the indirect solution method presented here boasts straightforward calculation steps and imposes minimal image format requirements. This research provides theoretical and technical support for blast-furnace slag's direct fiber forming process.

This study proposes an image recognition method to assist swine farm managers in collecting health data related to pig feeding and excretion. Analyzing the correlation between sow health data and indoor air quality in the pigsties revealed significant influences of air quality indicators on sow health. Increased levels of TVOC, CO2, and temperature were found to negatively affect feeding health, while increased temperature, humidity, and PM10 were found to negatively impact excretion health. These findings provide a basis for evaluating variables when constructing future sow disease prediction models. Analyzing the correlation between pig health status and air quality can help swine farm managers improve sow production environments, reduce disease risks, enhance production efficiency, and provide valuable insights for future research on disease prediction models.

Leukemia is a cancer of the bone marrow, a spongy tissue that secretes into the bones and serves as the site for the production of blood cells. One of the most prevalent kinds of leukemia in adults is acute myeloid leukemia (AML). Leukemia has non-specific signs and symptoms that are also similar to those of other interpersonal illnesses. The only way to accurately diagnose leukemia is by manually examining a stained blood smear or bone marrow aspirate under the microscope. However, this approach takes more time and is less precise. This paper describes a method for the automatic recognition and classification of AML in blood smears. Classification techniques include decision trees, logistic regression, support vector machines, and naive bayes.

Infrared and visible image fusion technologies are used to characterize the same scene by diverse modalities. However, most existing deep learning-based fusion methods are designed as symmetric networks, which ignore the differences between modal images and lead to the source image information loss during feature extraction. In this paper, we propose a new fusion framework for the different characteristics of infrared and visible images. Specifically, we design a dual-stream asymmetric network with two different feature extraction networks to extract infrared and visible feature maps respectively. The transformer architecture is introduced in the infrared feature extraction branch, which can force the network to focus on the local features of infrared images while still obtaining their contextual information. And the visible feature extraction branch uses residual dense blocks to fully extract the rich background and texture detail information of visible images. In this way, it can provide better infrared targets and visible details for the fused image. Experimental results on multiple datasets indicate that DSA-Net outperforms state-of-the-art methods in both qualitative and quantitative evaluations. In addition, we also apply the fusion results to the target detection task, which indirectly demonstrates the fusion performances of our method.

Loess is a special soil with high water sensitivity, which covers a large area in northwest China. Cracks are prone to generate in loess under the arid and semi-arid climate, which will provide the preferential channel for water and reduce the mechanical properties of soils. It is of great significance to understand the evolution characteristics and mechanisms of the cracks in loess areas. At present, the researches on cracks in soils are mainly concentrated in the characteristics of different cracking patterns. However, the spacing cracks are mostly discussed in materials like rock and concrete, rather than soils. The cracking characteristics and mechanisms of spacing cracks in loess is still inadequate. In this research, the drying tests of loess are carried out with different sizes of specimens. The parameters of spacing cracks and local strain distributions of the loess samples are obtained via PCAS and DIC methods during desiccation, respectively. The cracking modes, space cracking laws like insertion and saturation of spacing cracks are revealed. Finally, the size effect on the characteristics of spacing cracks is verified with a discrete element software MatDEM.

Zoom dysmorphia (ZD) is a facial dysmorphia triggered or aggravated by frequent virtual meetings. The frequent use of videoconferencing platforms has been linked to a distorted perception of facial images as individuals have an increased awareness of their appearance, given constant video feedback. As a result, dysmorphic concerns can develop. It is crucial to identify ZD as the condition interferes with the individual’s life and can trigger or aggravate body dysmorphic disorder (BDD). A standardized approach for screening ZD in non-psychiatric settings has yet to be defined. We discuss the features of ZD and the challenges of screening for ZD in a virtual setting. To facilitate the recognition of ZD in telehealth consultations, we propose a comprehensive ZD screening questionnaire that includes questions related to typical ZD features and a BDD-focused question. The questionnaire is concise and allows the identification of individuals with a potential ZD. A BDD assessment in such individuals should follow.

.Purpose: This paper presents a comparative analysis of state-of-the-art DIR algorithms in the context of head and neck squamous cell carcinoma (HNSCC) radiation therapy. Materials and Methods: We used a dataset of head and neck cancer patients within the Cancer Image Archive, which includes 31 patients with 3D CT image data from pre-treatment, mid-treatment, and post-treatment stages of radiotherapy. We applied DIR algorithms to two datasets, one for pre-treatment to mid-treatment registration and another for mid-treatment to post-treatment registration, by aligning one CT image to another CT image. To quantitatively analyze the patient data, we focused on the spinal cord, optic nerve, brain stem, cochlea, and PTV as the OAR datasets. To verify the DIR algorithm, we used two non-learning-based methods, SyN and NiftyReg, as well as three deep learning-based methods, Voxelmorph, Cyclemorph, and Transmorph. We trained a deep learning model by adjusting the ratio of the training dataset, validation dataset, and test dataset to 7:1:2. Results: The average DSCs of SyN and NiftyReg, the non-learning-based methods, were 0.74±0.06 and 0.70±0.12, respectively. The deep learning-based methods, Voxelmorph, Cyclemorph, and Transmorph, had average DSCs of 0.72±0.08, 0.68±0.11, and 0.69±0.13, respectively. The deep learning DIR algorithm produced transformed outputs of OAR/PTV in a shorter time than other models, including commercial and conventional mathematical algorithms (At Inference(sec/images), Voxelmorph: 0.36, Cyclemorph: 0.3, Transmorph: 5.1, SyN: 140, NiftyReg: 40.2). Conclusions: In conclusion, our study provides a comprehensive comparison of commercial DIR algorithms with traditional and deep learning-based DIR methods. Our results demonstrate that commercial DIR algorithms outperform traditional DIR methods in terms of accuracy and efficiency. However, we also show that deep learning-based DIR methods have the potential to achieve similar performance to commercial algorithms with proper training and optimization. Our findings suggest that both commercial and deep learning-based DIR methods have their respective advantages and limitations, and the choice of method should depend on the specific requirements of the clinical application. Overall, our study contributes to ongoing efforts to improve the accuracy and efficiency of DIR methods for better patient outcomes.

With the explosion of the generation, transmission and sharing of image data over the Internet and other unsecured networks, the need and significance for the development of novel image encryption algorithms is unprecedented. In this research work, we propose a novel framework for image encryption that is based on two hyperchaotic maps, utilized in conjunction with the single neuron model (SNM). The framework entails three successive stages, where in every stage, a substitution box (S-box) is applied, then XORing with an encryption key is carried out. The S-boxes and the encryption keys are generated from the numerical solutions of the hyperchaotic maps and the SNM. The performance of the proposed framework is gauged through a number of metrics, reflecting superior performance and complete asymmetry between the plain images and their encrypted versions. However, the main advantages of this work are: (1) The attained width of the key space; and (2) The achieved efficiency in software implementation. The superior key space of 2^{2551} is the result of employing the two hyperchaotic maps, while the improved efficiency, resulting in an average encryption rate of 8.54 Mbps, is the result of the utilization of the single neuron model, as well as the employment of optimized parallel processing techniques. In addition, the proposed encryption framework is shown to output encrypted images that pass the NIST SP 800 suite.

Accurate image registration is essential for eye movement compensation in optical coherence tomography (OCT) and OCT angiography (OCTA). The spatial resolution of an OCT instrument is typically anisotropic, i.e., different resolutions in the lateral and axial dimensions. When OCT images have anisotropic pixel resolution, residual distortion (RD) and false translation (FT) were always observed after image registration for rotational movement. In this study, RD and FT were quantitively analyzed over different degrees of rotational movement and various lateral and axial pixel resolution ratio (RL/RA) values. The theoretical analysis confirmed that the RD and FT increase significantly with the rotation degree and RL/RA. An image resizing assisting registration (RAR) strategy was proposed for accurate image registration. The performance of direct registration (DR) and RAR for retinal OCT and OCTA images were quantitatively compared. Experimental results confirmed that unnormalized RL/RA causes RD and FT, RAR can effectively improve the performance of OCT and OCTA image registration and distortion compensation.

The quality of yarn is essential in the control of the fabrics processes. There is some commercial equipment that measures the quality of yarn based on sensors, of different types, used for collecting data about some yarn characteristic parameters. The existing equipment, for the above-mentioned purpose, is characterized by its high size and cost, and for allowing the analysis of only few yarn quality parameters. Thus, this paper presents the development and results obtained with the design of a mechatronic prototype integrating a computer vision system that allows, among other parameters, the analysis and classification, in real time, of the hairs of the yarn using artificial intelligence techniques. The system also determines other characteristics inherent to the yarn quality analysis, such as: linear mass, diameter, volume, twist orientation, twist step, average mass deviation, coefficient of variation, hairiness coefficient, average hairiness deviation and standard hairiness deviation, as well as performs spectral analysis. A comparison of the obtained results with the designed system and a commercial equipment was performed validating the undertaken methodology.

Fringe projection profilometry(FPP) has been extensively applied in various fields for its superior fast speed, high accuracy and high data density. However, measuring some objects with high reflective surfaces or high dynamic range surfaces remains challenging for FPP. Some multiple exposure image fusion methods have been proposed and successfully improved the measurement performance for these kinds of objects. Normally, these methods have a relatively fixed sequence of exposure settings determined by artificial experience or trial and error experiments, which may decrease the efficiency of the entire measurement process and may have less robustness to various environmental lighting conditions and object reflective properties. In this paper, a novel self-adaptive multiple exposure image fusion method is proposed with two main aspects of improvement on adaptively optimizing the initial exposure and the exposure sequence. By introducing the theory of information entropy, combined with the analysis of the characterization of fringe image entropy, an adaptive initial exposure searching method is first proposed. Then, an exposure sequence generation method based on dichotomy is further described. On the base of these two improvements, a novel self-adaptive multiple exposure image fusion method for FPP as well as its detailed procedures are given. Experimental results validate the performance of the proposed self-adaptivity multiple exposure image fusion method by measuring the objects with different surface reflectivity and in different ambient lighting conditions.

This study presents the algorithm ACTION, defined as Automatic Center detection of Tropical cyclone (TC) using Image processing based on the Operational radar Network. Based on the high visibility of weather radar imagery, the TC’s motion vector is calculated from the continuous image change using optical flow, producing its rotation center as the TC’s center. The algorithm’s performance was verified for typhoons (TCs in the Northwestern Pacific) affecting the Korean Peninsula from 2018–2021 and showed a high detection rate of 80.8% within an error distance of 40 km compared to the best track of the Korea Meteorological Administration (KMA). The detection rate was 100% for typhoons with temporally consistent morphological characteristics. ACTION automatically generates TC center information upon the TC’s initial entry inside the observation radius even in the absence of uniform radar data. ACTION easily calculates using Open Source Computer Vision, performs in real time, and can be directly applied to rapidly generated weather radar images; hence, it is currently being utilized by the KMA. The high-temporal-resolution TC center information calculated through ACTION is expected to improve the efficiency of TC forecasting.

Numerous new technologies have been implemented in image analysis methods that help researchers withdraw scientific conclusions from biological phenomena. Plants of the family Lemnaceae (duckweeds) are the smallest flowering plants in the world, and biometric measurements of single plants and their growth rate are highly challenging. Although the use of software for digital image analysis has changed the way scientists extract phenomenological data (also for studies on duckweeds), the procedure is often not wholly automated and sometimes relies on the intervention of a human operator. Such a constraint can limit the objectivity of the measurements and generally slows down the time required to produce scientific data. Here is the need to implement image analysis software with artificial intelligence that can substitute the human operator. In this paper, we present a new method to study the growth rates of the plants of the Lemnaceae family based on the application of machine learning procedures to digital image analysis. The method is compared to existing analogical and computer-operated procedures. Results showed that our method drastically reduces the time consumption of the human operator while retaining a high correlation in the growth rates measured with other procedures As expected, machine learning methods applied to digital image analysis can overtake the constraints of measuring growth rates of very small plants and might help duckweeds gain worldwide attention thanks to their great nutritional qualities and biological plasticity.

Nonlinear catastrophes caused by geological fluids are a fundamental issue in rock mechanics and the geoengineering hazard field. For the consideration of hydrodynamic force on red-bed mudstone softening damage, X-ray visualization test on the fissure flow in mudstone block failure under hydrodynamic force was performed in this study based on block scale and the physical phenomena of fissure seepage and nonlinear diffusion were further explored. A new method for evaluating the hydro-damage degrees of rocks using an X-ray image analysis was proposed, and the quantitative relation of diffusion coefficients of hydro-damage and seepage was established. The research results revealed that the hydrodynamic force promoted the fluid-filled fissure behavior in mudstone specimen failure. Also, the seepage and diffusion phenomena of fluid in rocks during failures were indicated using X-ray imaging. A dual mechanical behavior was presented in the nonlinear seepage and abnormal diffusion of a red mudstone geological body under hydrodynamic conditions. The damaged degree of mudstone was aggravated by the effect of hydrodynamic force, and the initial seepage–diffusion coefficient with respect to lower hydro-damage was larger than the final seepage–diffusion coefficient with respect to higher hydro-damage of rocks with a decreasing nonlinear trend.

Cellular Nonlinear Networks (CNN) are a concept introduced in 1988 by Leon Chua and Lin Yang as a bio-inspired architecture, capable of massively parallel computation. Later on, CNN have been enhanced by incorporating designs that incorporate memristors to profit from their processing and memory capabilities. In addition, Stochastic Computing (SC) can be used to optimize the quantity of required processing elements; thus it provides a lightweight approximate computing framework, quite accurate and effective, though. In this work, we propose utilization of SC in designing and implementing a memristor-based CNN. As a proof of the proposed concept, an example of application is presented. This application combines Matlab and a FPGA in order to create the CNN. The implemented CNN has then been used to perform three different real-time applications on a 512x512 gray-scale and a 768x512 color image: storage of the image, edge detection, and image sharpening. It has to be pointed out that the same CNN has been used for the three different tasks, with the sole change of some programmable parameters. Results show an excellent capability with significant accompanying advantages, like the low number of needed elements further allowing for a low cost FPGA-based system implementation, something confirming the system’s ability for real time operation.

Gradient-domain image processing is a technique where, instead of operating directly on the image pixel values, the gradient of the image is computed and processed. The resulting image is obtained by reintegrating the processed gradient. This is normally done by solving the Poisson equation, most oftenly by means of a finite difference implementation of the gradient descent method. However, this technique in some cases lead to severe haloing artefacts in the resulting image. To deal with this, local or anisotropic diffusion has been added as an ad-hoc modification of the Poisson equation. In this paper, we show that a version of anisotropic gradient-domain image processing can result from a more general variational formulation through the minimisation of a functional formulated in terms of the eigenvalues of the structure tensor of the differences between the processed gradient and the gradient of the original image. Example applications of linear and non-linear local contrast enhancement and colour image daltonisation illustrate the behaviour of the method.

Sediment plumes, released to the Bering Sea from the delta front of the Yukon River, Alaska, are initiated mainly by glacier-melt sediment runoffs in the glacierized regions of the Yukon River drainage basin. The surface sediment plumes are extended around the fan-shaped Yukon River delta, which is followed by the northwestward dispersion. During continuous measure-ments of the Yukon River discharge and sediment load, behaviors of the sediment plumes were explored by shipboard observations in the Bering Sea offshore from the Yukon delta. At the high river sediment load of ca. 3000 kg/s, the plume partially plunged into the sea bottom layer. The plunging probably originated in the nepheloid-layer formation from the flocculation of river-suspended sediment, of which more than 90 %wt. is silt and clay (grain size d < 0.063 mm). In order to numerically obtain the area of the surface sediment plumes, a satellite image analy-sis was performed by using three near-infrared bands in MODIS/Aqua or MODIS/Terra. The plume area was significantly correlated (R2=0.735, p<0.01) to the sediment load averaged for the two days with time lags of 20 days and 21 days to the date of a certain satellite image. Hence, the dispersion of plume-suspended sediment appears to be controlled by the sediment runoff events in the Yukon River rather than the northward “Alaskan Coastal Water”.

A progressive FEA fatigue degradation model for composites was developed and implemented using a UMAT user material subroutine in Abaqus. Numerical results were compared to experimental strain field data from high frequency Digital Image Correlation (DIC) of split disk fatigue testing of pressure vessel cut outs with holes. The model correctly predicted the onset and evolution of damage in the matrix as well as the onset of fiber failure. The model use progressive failure analysis based on the maximum strain failure criterion, the cycle jump method and Miner sum damage accumulation rule. A parameter study on matrix properties was needed to capture the scatter in strain fields observed experimentally by DIC. S-N curve for the matrix material had to be lowered by 0% to 60% to capture the experimental scatter. The onset of local fiber failure had to be described by local S-N curves measured by DIC having 2.5 times greater strain than that of S-N curves found from standard coupon testing.

Existing methods enhance mid-low resolution remote sensing ship detection by feeding super-resolved images to the detectors. Although these methods marginally improve the detection accuracy, the correlation between image super-resolution (SR) and ship detection is under-exploited. In this paper, we propose a simple but effective ship detection method called ShipSR-Det, in which both the output image and the intermediate features of the SR module are fed to the detection module. Using the super-resolved feature representation, the potential benefit introduced by image SR can be fully used for ship detection. We apply our method to the SSD and Faster-RCNN detectors and develop ShipSR-SSD and ShipSR-Faster-RCNN, respectively. Extensive ablation studies validate the effectiveness and generality of our method. Moreover, we compare ShipSR-Faster-RCNN with several state-of-the-art ship detection methods. Comparative results on the HRSC2016, DOTA and NWPU VHR-10 datasets demonstrate the superior performance of our proposed method.

Modernist planners were given the opportunity to apply the idea of ​​the Modern Movement in the implementation of cities transforming after World War II, which was not possible before in long-term evolving structures. Usually, however, it was impossible to change everything: some urban and even architectural elements were necessary to be preserved. As a result, the elements of the ancient landscapes of these cities - historical objects and spaces - necessary for the continuity of evidence of history and identity have been preserved, sometimes through reconstruction. Finally, in some of these cities, both old and new elements contribute to the contemporary urban landscape. The aim of the research is to examine two key categories of urban structure in this perspective: 1. landmarks 2. urban network nodes. For this purpose: 1. the changes in the number and rank of the space with which the landmarks exhibition was carried out were compared; 2. the changes in the rank of nodes in the urban structure functioning both in the pre-war and in the present structure, as well as changes in their distribution and concentration areas were analyzed. The numbers of individual elements were balanced and the changes in the structure of their arrangement were analyzed with the use of a polygon grid. The analyzes were based on the case of a medium-scale European city (Białystok). Research shows that as a result of post-war reconstruction, the number of elements of the urban grid decreased, especially those of lower rank, the rank of others decreased.

In the last years, more and more studies highlight the advantages of complementing traditional master classes with additional activities that improve students´ learning experience. This combination of teaching techniques is specially advised in the field of structural engineering, where intuition of the structural response it is of vital importance to understand the studied concepts. This paper deals with the introduction of a new (and more encouraging) educational tool to introduce intuitively students in the dynamic response of structures excited with an educational shaking table. Most of the educational structural health monitoring systems use sensors to determine the dynamic response of the structure. The proposed tool is based on a radically different approach, as it is based on low-cost image-recognition techniques. In fact, it only requires the use an amateur camera, a black background and a computer. In this study, the effects of both the camera location and the image quality are also evaluated. Finally, to validate the applicability of the proposed methodology, the dynamic response of small-scale buildings with different typologies is analyzed. In addition, a series of surveys were conducted in order to evaluate the activity based on student´s satisfaction and the actual acquisition and strengthening of knowledge.

Defensive structure is important in transportation field for kinds of intentional or unintentional explosion. Structures subjected to unconfined and confined explosion will bear different blast loads and their dynamic responses are different. The present work focus on the dynamic response mechanism of steel plate under unconfined and confined blast loads through both numerical and experimental studies. In the experiment, Digital Image Correlation (DIC) technique was applied to record and analyze the dynamic response process of large-scale field blast test. The DIC measured curve and the numerical calculated curves agrees well in both the trends and the peak values. Then, the dynamic response mechanism of steel plate under unconfined blast (UB) load and confined blast (CB) load were compared and discussed. Results show that the dynamic response of plate can be divided into three phases under both UB and CB loads, while with different mechanism. In phase I, plastic hinge starts from the center and moves to the boundary in UB condition, while in case of CB, plastic hinge occurs close to the boundary and moves in the opposite direction. In phase II, two plastic hinge lines propagates towards each other, and a platform exists between the boundary and the central area remains undeformed in UB condition, while in CB condition, larger deformation in peripheral region rather than central area produces.

A lightweight image encryption algorithm based on chaos induction via a 5-dimensional hyperjerk oscillator (5DHO) is presented. First, the dynamics of our 5DHO network is investigated and shown to exhibit up to five coexisting hidden attractors in the state space that depend exclusively on the system’s initial values. Further, a simple implementation of the circuit was used to validate its ability to exhibit chaotic dynamical properties. Second, an Arduino UNO platform is used to confirm the usability of our oscillator in digital implementation of the system. Finally, an efficient image encryption application is executed using the proposed chaotic networks based on the use of permutation-substitution sequences. The superior qualities of the proposed strategy are traced to the dynamic set of keys used in the substitution process which heralds the generation of the final ciphered image. Based on the results obtained from the entropy analysis (7.9976), NPCR values (99.62), UACI tests (33.69) and encryption execution time for 512x512 images (0.1179 sec), the proposed algorithm is adjudged to be fast and robust to differential and statistical attacks relative to similar approaches.

Electron Beam Melting (EBM) is an increasingly used Additive Manufacturing (AM) technique employed by many industrial sectors, including the medical device and aerospace industries. In EBM process monitoring, data analysis for processed layer quality evaluation is currently focused on the extraction of information from the raw data collected in-EBM process, i.e. thermal/ optical / electronic images, and the comparison between the collected data and the Computed Tomography (CT)/ microscopy images generated post-EBM process. This article postulates that a stack of bitmaps could be generated from the 3D model at a range of Z heights during file preparation of the EBM process, and serve as a reference image set. In-EBM process comparison between that and the workpiece images collected during the EBM process could then be used for quality assessment purposes. In addition, despite the extensive literature on 3D model slicing and contour generation for AM process preparation, no methods regarding image generation from cross sections of the 3D models have been disseminated in details. This article aims to address this by presenting a piece of 3D model-image generation software. The software is capable of generating binary 3D model reference images with user-defined Region-of-Interest (ROI) of the processing area, and Z heights of the model. It is envisaged that this 3D model-reference image generation ability opens up new opportunities in quality assessment for the in-process monitoring of the EBM process.

Automated skin lesion analysis is one of the trending fields that has gained attention among the dermatologists and healthcare practitioners. Skin lesion restoration is an essential preprocessing step for lesion enhancements for accurate automated analysis and diagnosis. Digital hair removal is a non-invasive method for image enhancement by solving the hair-occlusion artefact in previously captured images. Several hair removal methods were proposed for skin delineation and removal. However, manual annotation is one of the main challenges that hinder the validation of these proposed methods on a large number of images or using benchmarking datasets for comparison purposes. In the presented work, we propose a realistic hair simulator based on context-aware image synthesis using image-to-image translation techniques via conditional adversarial generative networks for generation of different hair occlusions in skin images, along with the ground-truth mask for hair location. Besides, we explored using three loss functions including L1-norm, L2-norm and structural similarity index (SSIM) to maximise the synthesis quality. For quantitatively evaluate the realism of image synthesis, the t-SNE feature mapping and Bland-Altman test are employed as objective metrics. Experimental results show the superior performance of our proposed method compared to previous methods for hair synthesis with plausible colours and preserving the integrity of the lesion texture.

Image processing is a field of which its popularity increases and continues and, that grows dynamically with new technologies. Nowadays, image processing finds itself in use in almost every field. One of these uses is undoubtedly in the field of medicine, where diagnosis and treatment planning are made from images and, which is constantly changing with newly developed techniques. Of course, the most important factor in using this so widely in the medical field is the acquisition of images on every medical field. With the help of these images, the complaints can be seen more easily and the doctor can follow a path in the treatment of the disease. In our study, we used the Region Growing segmentation method to detect the intestinal mass. This study compares the area determined by the specialist with the area obtained with the segmentation process and, it is seen that the created software system can be used as an auxiliary system to specialist doctors.

Capsule endoscopy, which uses a wireless camera to take images of the digestive tract, is emerging as an alternative to traditional colonoscopy. The diagnostic values of these images depend on the quality of revealed underlying tissue surfaces. In this paper, we consider the problem of enhancing the visibility of detail and shadowed tissue surfaces for capsule endoscopy images. Using concentric circles at each pixel for random walks combined with stochastic sampling, the proposed method enhances the details of vessel and tissue surfaces. The framework decomposes the image into two detail layers that contain shadowed tissue surfaces and detail features. The target pixel value is recalculated for the smooth layer using similarity of the target pixel to neighboring pixels by weighting against the total gradient variation and intensity differences. In order to evaluate the diagnostic image quality of the proposed method, we used clinical subjective evaluation with a rank order on selected KID image database and compared to state of the art enhancement methods. The result showed that the proposed method provides a better result in terms of diagnostic image quality and objective quality contrast metrics and structural similarity index.

Grading of rice grains has gain attentions due its requirement of quality assessment during import or export. Rice grain quality depends on milling operation, where rice hull is removed with a huller system followed by whitening operation. In such process, adjustment of rollers, control, and operation is important in terms of quality of milled rice. Especially, the basmati rice needed more quality assurance as it is not parboiled rice and exported globally with a high product value. In this present work, the basic problem of quality assessment in rice industry is addressed with digital image processing based technique. Machine vision and digital image processing provide an alternative with the automated, nondestructive, cost-effective, and fast approach as compared with traditional method which is done manually by human inspectors. A model of quality grade testing and identification is built based on morphological features using digital image processing and knowledge based adaptive neuro-fuzzy inference system (ANFIS). The qualities of rice kernels are determined with the help of shape descriptors and geometric features using the sample images of milled rice. The adopted technique has been tested on a sufficient number of training images of basmati rice grain. The proposed method gives a promising result in an evaluation of rice quality with 100% classification accuracy for broken and whole grain. The milling efficiency is also assessed using the ratio between head rice and broken rice percentage and it is 77.27% for the test sample. The overall results of the adopted methodology are promising in terms of classification accuracy and efficiency.

Direct georeferencing of airborne pushbroom scanner data usually suffers from the limited precision of navigation sensors onboard of the aircraft. The bundle adjustment of images and orientation parameters, used to perform geocorrection of frame images during the post-processing phase, cannot be used for pushbroom cameras without difficulties: it relies on matching corresponding points between scan lines, which is not feasible in the absence of sufficient overlap and texture information. We address this georeferencing problem by equipping our aircraft with both a frame camera and a pushbroom scanner: the frame images and the navigation parameters measured by a couple GPS/Inertial Measurement Unit (IMU) are input to a bundle adjustment algorithm; the output orientation parameters are used to project the scan lines on a Digital Elevation Model (DEM) and on an orthophoto generated during the bundle adjustement step; using the image feature matching algorithm Speeded Up Robust Features (SURF), corresponding points between the image formed by the projected scan lines and the orthophoto are matched, and through a least-squares method, the boresight between the two cameras is estimated and included in the calculation of the projection; finally, using Particle Image Velocimetry (PIV) on the gradient image, the projection is deformed into a final image that fits the geometry of the orthophoto. We apply this algorithm to five test acquisitions over Lake Geneva region (Switzerland) and Lake Baikal region (Russia). The results are quantified in terms of Root Mean Square Error (RMSE) between matching points of the RGB orthophoto and the pushbroom projection. From a first projection where the Interior Orientation Parameters (IOP) are known with limited precision and the RMSE goes up to 41 pixels, our geocorrection estimates IOP, boresight and Exterior Orientation Parameters (EOP) and produces a new projection which RMSE with the reference orthophoto is around two pixels.

The IoT Internet of Things being a promising technology of the future. It is expected to connect billions of devices. The increased communication number is expected to generate data mountain and the data security can be a threat. The devices in the architecture are fundamentally smaller in size and low powered. In general, classical encryption algorithms are computationally expensive and this due to their complexity and needs numerous rounds for encrypting, basically wasting the constrained energy of the gadgets. Less complex algorithm, though, may compromise the desired integrity. In this paper we apply a lightweight encryption algorithm named as Secure IoT (SIT) to a quantized speech image for Secure IoT. It is a 64-bit block cipher and requires 64-bit key to encrypt the data. This quantized speech image is constructed by first quantizing a speech signal and then splitting the quantized signal into frames. Then each of these frames is transposed for obtaining the different columns of this quantized speech image. Simulations result shows the algorithm provides substantial security in just five encryption rounds.

Olive trees have been of economic and cultural value since pre-Roman times, and continue to dominate landscapes and agriculture in many mediterranean regions. Recent mass losses of olive trees in Southern Italy due to an exotic plant pathogen highlight the need for methods that to monitor the olive trees in a landscape or region operationally. Here, we develop a method for counting olive trees from aerial photographs and test it in areas with a high diversity of olive tree ages, sizes, and shapes. This heterogeneity complicates tree counting as centennial trees often have crowns that are split into multiple segments, resembling multiple crowns, while nearby crowns often form a semi-closed canopy comprising multiple trees. Comparisons with reference counts in two 20 ha sites and over three different years indicate the automated counts tend to be reasonably accurate (median error 13%, n = 6), but heavily influenced by a few olive orchards with particularly high planting densities and a relatively closed canopy in which distinguishing individual trees is challenging. Overall, the algorithm estimated tree densities well (counting 82 to 109 trees/ha versus 87 to 104 trees/ha in the reference counts), indicating the method is suitable to track the number of olive trees over large areas.

With the development of hyperspectral technology, to establish an effective spectral data compressive reconstruction method that can improving data storage, transmission and maintaining spectral information is critical for quantitative remote sensing research and application in vegetation. By introducing the idea of compressive sensing in compressed reconstruction, the spectral adaptive grouping distributed compressive sensing algorithm is proposed, which enables a distributed compressed sensing reconstruction of plant hyperspectral data. The experimental results showed that comparing with orthogonal matching pursuit(OMP) and gradient projection reconstruction(GPSR), the proposed algorithm can significantly improve the visual effect of image reconstruction in the spatial domain. The PSNR in low sampling rate(sampling rate is lower than 0.2) increases by 13.72dB than OMP and 1.66dB than GPSR. In the spectral domain, the average normalized root mean square error、the mean absolute percentage error and the mean absolute error of the proposed algorithm is35.38%，31.83% and 33.33% lower than GPSR respectively.. Therefore, the proposed algorithm can achieve relatively high reconstructed efficiency.

This study aims to test a model linking destination image, perceived value, tourist satisfaction, and tourist loyalty. Based on a sample of 300 tourists travelling by car from the World Natural Heritage Site of Tianchi, China, a new model of destination image was explored and data were analysed using partial least squares structural equation modelling (PLS-SEM). The results show that perceived value and satisfaction are direct antecedents of destination loyalty. Above all, perceived value and tourist satisfaction mediate the relationship between destination image and loyalty. Finally, this study discusses the theoretical and management implications of the findings in order to boost the tourism industry in the context of car trips.

For Visually impaired People (VIPs), the ability to convert text to sound can mean a new level of independence or the simple joy of a good book. With significant advances in Optical Character Recognition (OCR) in recent years, a number of reading aids are appearing on the market. These reading aids convert images captured by a camera to text which can then be read aloud. However, all of these reading aids suffer from a key issue – the user must be able to visually target the text and capture an image of sufficient quality for the OCR algorithm to function – no small task for VIPs. In this work, a Sound-Emitting Document Image Quality Assessment metric (SEDIQA) is proposed which allows the user to hear the quality of the text image and automatically captures the best image for OCR accuracy. This work also includes testing of OCR performance against image degradations, to identify the most significant contributors to accuracy reduction. The proposed No-Reference Image Quality Assessor (NR-IQA) is validated alongside established NR-IQAs and this work includes insights into the performance of these NR-IQAs on document images.

The production of cheese, a beloved culinary delight worldwide, faces challenges in maintaining consistent product quality and operational efficiency. One crucial stage in this process is determining the precise cutting time during curd formation, which significantly impacts the quality of the cheese. Misjudging this timing can lead to the production of inferior products, harming a company's reputation and revenue. Conventional methods often fall short of accurately assessing variations in coagulation conditions due to the inherent potential for human error. To address this issue, we propose an anomaly detection-based approach. In this approach, we treat the class representing curd formation as the anomaly to be identified. Our proposed solution involves utilizing a one-class, fully convolutional data description network, which we compared against several state-of-the-art methods to detect deviations from the standard coagulation patterns. Encouragingly, our results show F1 scores of up to 0.92, indicating the effectiveness of our approach.

We propose a novel unsupervised semantic segmentation method for fast and accurate flood area detection utilizing color images acquired from Unmanned Aerial Vehicles (UAVs). To the best of our knowledge, this is the first fully unsupervised method for flood area segmentation in color images captured by UAVs, without the need of pre-disaster images. The proposed framework addresses the problem of flood segmentation based on parameter-free calculated masks and unsupervised image analysis techniques. First, a fully unsupervised algorithm gradually excludes areas classified as non-flood utilizing calculated masks over each component of the LAB colorspace, as well an RGB vegetation index and the detected edges of the original image. Unsupervised image analysis techniques, such as distance transform, are then applied, producing a probability map for the location of flooded areas. Finally, flood detection is obtained by applying the hysteresis thresholding segmentation. The proposed method is tested and compared with variations, and other supervised methods in two public datasets, consisting of 953 color images in total, yielding high-performance results, with 87.4% and 80.9% overall accuracy and F1-Score, respectively. The results and computational efficiency of the proposed method show that it is suitable for on board data execution and decision-making during UAVs flight.

The essence of cross-modal generation and retrieval modeling for image-text interactions lies in semantic consistency. Among these, cross-modal representation alignment serves as the foundational requirement for achieving cross-modal semantic consistency. The crux of our proposed method entails the collaborative training of an image captioning model(referred to as the ’captioner’) and a pair of contrastive encoders for image-text matching(referred to as the ’concoder’). This synergistic approach yields concurrent enhancements in concoder’s performance, captioner’s proficiency, and cross-modal semantic coherence. We coin the proposed method as ‘ReCap’. To begin, we initialize the concoder through knowledge distillation from a pre-trained model. During the training process, we use the output of the initialized concoder as input to a residual attention network, jointly training it with the captioner to achieve semantic consistency. Subsequently, we iteratively update the concoder during the image-text momentum encoding phase using the residual attention network, creating a closed-loop for semantic consistency resolution. Upon completion of training, the concoder and captioner modules can be used independently to perform tasks related to cross-modal retrieval and cross-modal generation of text and images. In order to substantiate the efficacy of our proposed method, we initially conducted empirical experiments encompassing image-text retrieval and image captioning tasks, employing a widely recognized benchmark dataset, MSCOCO. Subsequently, we assessed the fine-grained cross-modal alignment performance of the concoder through an image classification task on the iNaturalist 2018 dataset. The achieved performance metrics notably surpassed those achieved by several incumbent state-of-the-art models, thus validating the proficiency of our method. Finally, motivated by the practical requirements of handling the nature conservation image data, we performed caption annotation for the iNaturalist 2018 dataset. Subsequently, we trained the ReCap on this annotated data. Experimental results demonstrate that our proposed method can maintain semantic consistency between cross-modal retrieval and image caption generation for species with similar visual features but distinct semantics. This achievement signifies a meaningful exploration in the field of cross-modal semantic consistency representation, holding significant practical value in the domain of biological research.

Verifying the accuracy of a 3D-printed object involves using an image processing system. This system compares images of the CAD model of the object to be printed with its 3D-printed counterparts to identify any discrepancies. It is important to note that the integrity of the accuracy measurement is heavily dependent on the image processing settings chosen. This study focuses on this issue by developing a customized image processing system. The system generates binary images of a given CAD model and its 3D-printed counterparts and then compares them pixel-by-pixel to determine the accuracy. Users can experiment with various image processing settings, such as grayscale to binary image conversion threshold, noise reduction parameters, masking parameters, and pixel-fineness adjustment parameters, to see how they affect accuracy. The study concludes that the grayscale to binary image conversion threshold has the most significant impact on accuracy and that the optimal threshold varies depending on the color of the 3D-printed object. The system can also effectively eliminate noise (filament marks) during image processing, ensuring accurate measurements. Additionally, the system can measure the accuracy of highly complex porous structures where the pores size, depth, and distribution are random. The insights gained from this study can be used to develop intelligent systems for the metrology of additive manufacturing.

Underwater image often exhibit detail blurring and color distortion due to light scattering, impurities, and other influences, obscuring essential textures and details. This presents a challenge for existing super-resolution techniques in identifying and extracting effective features, making high-quality reconstruction difficult. This research aims to innovate underwater image super-resolution technology to tackle this challenge. Initially, an underwater image degradation model was created by integrating random subsampling, Gaussian blur, mixed noise, and suspended particle simulation to generate a highly realistic synthetic dataset, thereby training the network to adapt to various degradation factors. Subsequently, to enhance the network's capability to extract key features, improvements were made based on the symmetrically structured Blind Super-Resolution Generative Adversarial Network (BSRGAN) model architecture. An attention mechanism based on energy functions was introduced within the generator to assess the importance of each pixel, and a weighted fusion strategy of adversarial loss, reconstruction loss, and perceptual loss was utilized to improve the quality of image reconstruction. Experimental results demonstrate that the proposed method achieved significant improvements in Peak Signal-to-Noise Ratio (PSNR) and Underwater Image Quality Measure (UIQM) by 0.85 dB and 0.19, respectively, significantly enhancing the visual perception quality and indicating its feasibility in super-resolution applications.

Diabetic retinopathy (DR) is the primary factor leading to vision impairment and blindness in diabetic people. Uncontrolled diabetes can damage the retinal blood vessels. Initial detection and prompt medical intervention are vital in preventing progressive vision impairment. Today's growing medical field poses more significant workload and diagnostic demands on medical professionals. In the proposed study, a convolutional neural network (CNN) is employed to detect the stages of DR. This research is crucial for studying DR because of its innovative methodology incorporating two different public datasets. This strategy enhances the model's capacity to generalize unseen DR images, as each dataset encompasses unique demographics and clinical circumstances. The network can learn and capture complicated hierarchical image features with asymmetric weights. Each image undergoes preprocessing using contrast-limited adaptive histogram equalization (CLAHE) and the discrete wavelet transform (DWT). The model is trained and validated using the combined datasets of Dataset for Diabetic Retinopathy (DDR) and the Asia Pacific Tele-Ophthalmology Society (APTOS). The CNN model is tuned with different learning rates and optimizers. The CNN model with Adam optimizer achieved 70% accuracy and an area under curve (AUC) score of 0.82. The recommended study results may reduce diabetes-related vision impairment by early DR severity identification.

Over the last decades, an important number of regional and global Digital Elevation Models (DEMs) have been released publicly. As a consequence, researchers need to choose between several of these models to perform their studies and to use these DEMs as third-party data to compute derived products (e.g., for orthorectification). However, the comparison of DEMs is not trivial. For most quantitative comparisons, DEMs require to be expressed in the same Coordinate Reference System (CRS), sampled over the same grid (i.e., be at the same ground sampling distance with the same pixel-is-area or pixel-is-point convention), and with heights relative to the same Vertical Reference System (VRS). Thankfully, many open tools allow us to perform these transformations precisely and easily. Despite these rigorous transformations, local or global planimetric displacements may still be observed from one DEM to another. These displacements, or disparities, may lead to significant biases in comparisons of DEM elevations or derived products such as slope, aspect or curvature. Therefore, the control of DEM planimetric accuracy is certainly the most important task to perform at first before any comparison. This paper presents the disparity analysis method enhanced to achieve a sub-pixel accuracy, by interpolating the linear regression coefficients computed within an exploration window. This new method is significantly faster than oversampling the input data because it uses the correlation coefficients that have been already computed in the disparity analysis. To demonstrate the robustness of this algorithm, artificial displacements have been performed in a 11x11 grid with a 0.1-pixel step in both directions. The Copernicus DEM GLO-30 has been interpolated by a “classical” bicubic to achieve these 11x11 displacements. This validation method has been applied in four 10 km x 10 km DEMIX tiles showing different height distributions. Globally, this new sub-pixel accuracy method is robust. Artificial displacements have been retrieved with with typical errors (e_b) ranging from 12 to 20 % of the pixel size (worst case in Croatia).These errors in displacement retrievals are not constant in the 11x11 grid and the overall error E_b depends on the roughness (altimetry variations) encountered in the different tiles. The second idea of this paper is to assess the impact of the bicubic parameter (slope of the weight function at a distance d=1 of the interpolated point) on the accuracy of the displacement retrieval. By considering E_b as a quality indicator, tests have been performed in the four DEMIX tiles making the bicubic parameter vary between -1.5 and 0.0 by step of 0.1. For each DEMIX tile, the best bicubic (BBC) parameter b* is interpolated from the four E_b minimal values. This BBC parameter b* is low for flat areas (around -0.95) and higher in mountainous areas (around -0.75). The roughness indicator is the standard deviation of the slope norms computed from all the pixels of a tile. A logarithmic regression analysis performed between the roughness indicator and the BBC parameter b* computed in 67 DEMIX tiles shows a high correlation (r=0.717). The logarithmic regression formula estimating the BBC parameter from the roughness indicator is generic and may be applied to estimate the displacements between two different DEMs. This formula may also be used to set up a future Adaptative Best BiCubic (ABBC) that will estimate the local roughness in a sliding window to compute a local BBC.

Domain generalization has led to remarkable achievements in Hyperspectral Image (HSI) classification. Inspired by contrastive language-image pre-training (CLIP), the language-aware domain generalization method has been explored to learn cross-domain-invariant representation. However, existing methods face some challenges: 1) The weak capacity to extract long-range contextual information and inter-class correlation. 2) Due to the inadequacies of the large-scale pre-training for HSI data, the spatial-spectral features of HSI and linguistic features can not be straightforwardly alignment. To address the above problems, a novel network has been proposed with a CLIP framework, which consists of an image encoder, based on an encoder-only transformer to obtain the global contextual information and inter-class correlation, a frozen text encoder, and a cross-attention mechanism, named Linguistic-Interact-with-Visual Engager (LIVE), enhances the interaction between two modalities. Extensive experiments demonstrating superior performance over state-of-the-art methods in HSI Domain Generalization with a CLIP framework.

In this paper we develop and explore a methodical pipeline for swimmer localisation in outdoor environments. The developed framework is intended to be used for enhancing swimmer safety. A main issue we deal with by the proposed approach is the lack of real world training data in such outdoor environments. Natural lighting changes, dynamic water textures and possibly barely visible swimming persons are key elements to approach. We account for these difficulties by adopting an effective background removal technique with available training data. This allows us to edit swimmers into natural environment backgrounds for the use in subsequent image augmentation. We created 17 training datasets with real images, synthetic images and a mixture of both to investigate different aspects and characteristics of the proposed approach. The datasets are used to train a YOLO architecture for the possible future application in real-time detection. The trained framework is then tested and evaluated on outdoor environment imagery acquired by a safety drone to investigate and confirm the usefulness for outdoor swimmer localisation.

To accurately and non-destructively monitor the growth of spring wheat in the Hetao Irrigation Area, UAV remote sensing was employed during various fertility stages. Digital image indices from diverse fertilization treatments were calculated and compared with physiological indices to identify the most sensitive digital image indices corresponding to these indices. The study underscored the critical importance of the flowering stage in the growth of spring wheat, thus highlighting the necessity of focusing on this stage. This finding reiterated that the flowering stage was pivotal for spring wheat development in the Hetao Irrigation Area. Several digital image indices, such as GLA, R, G, INT, g, GRVI, MGRVI, RGBVI, EXG, and GRRI, exhibited a high frequency of significant correlations with physiological indices during the four primary reproductive stages of wheat. Consequently, these sensitive digital image indices during the flowering stage can more effectively characterize the physiological indices of spring wheat.

When it comes to financial transactions and counterfeit detection systems, banknote classification is crucial. In this paper, we propose an approach for the classification of Kazakhstan banknote images integrating Learning Vector Quantization (LVQ) with Statistical Texture Feature Extraction using the Gray Level Co-occurrence matrix (GLCM). Our methodology demonstrates effectiveness in accurately classifying banknote images, as evidenced by experimental results. The comprehensive testing scenarios show promising outcomes, with the combination of GLCM and Color Histogram under the LVQ algorithm achieving a high accuracy of 94.87 percent at distance 1 and 90°. These findings indicate the robustness and practical viability of the proposed method for authenticating banknotes.

Neural radiance fields (NeRF) leverage a neural representation to encode scenes, obtaining photo-realistic rendering of novel views. However, NeRF has notable limitations. A significant drawback is that it does not capture surface geometry and only renders the object surface colors. Furthermore, the training of NeRF is exceedingly time-consuming. We propose Depth-NeRF as a solution to these issues. Specifically, our approach employs a fast depth completion algorithm to denoise and complete the depth maps generated by RGB-D cameras. These improved depth maps guide the sampling points of NeRF to be distributed closer to the scene's surface, benefiting from dense depth information. Furthermore, we have optimized the network structure of NeRF and integrated depth information to constrain the optimization process, ensuring that the termination distribution of the ray is consistent with the scene’s geometry. Compared to NeRF, our method accelerates the training speed by 18% and significantly reduces the RMSE between the rendered scene depth and the ground truth depth, which indicates that our method can better capture the geometric information of the scene. With these improvements, we can train the NeRF model more efficiently and achieve more accurate rendering results.

People are drawn to woodcut-style designs due to their striking visual impact and strong contrast. However, traditional woodcut prints and previous computer-aided methods have not addressed the issues of dwindling design inspiration, lengthy production times, and complex adjustment procedures. We propose a novel network framework, the Woodcut-style Design Assistant Network (WDANet), to tackle these challenges. Notably, our research is the first to utilize diffusion models to streamline the woodcut-style design process. We've curated the Woodcut-62 dataset, featuring works from 62 renowned historical artists, to train WDANet in absorbing and learning the aesthetic nuances of woodcut prints, offering users a wealth of design references. Based on a noise reduction network, our dual cross-attention mechanism effectively integrates text and woodcut-style image features. This allows users to input or slightly modify a text description to quickly generate accurate, high-quality woodcut-style designs, saving time and offering flexibility. As confirmed by user studies, quantitative and qualitative analyses show that WDANet outperforms the current state-of-the-art in generating woodcut-style images and proves its value as a design aid.

Drones have been used in a variety of scenarios such as atmospheric monitoring, fire rescue, agricultural irrigation, etc., in which accurate environmental perception is of crucial importance for both decision-making and control. Among the drone sensors, the RGB camera is indispensable for capturing rich visual information for vehicle navigation but encounters a grand challenge in high-dynamic-range scenes that occur frequently in real applications. Specifically, the recorded frames suffer from under-exposure and over-exposure simultaneously and degenerate the successive vision tasks. To solve the problem, we take object tracking as an example and leverage the superior response of event cameras over a large intensity range to propose an event-assisted object tracking algorithm that can achieve reliable tracking under large intensity variations. Specifically, we propose to pursue feature matching from dense event signals, and based on which to (i) design a UNet-based image enhancement algorithm to balance the RGB intensity with the help of neighboring frames in the time domain, and then (ii) construct a dual-input tracking model to track the moving objects from intensity balanced RGB video and event sequence. The proposed approach is comprehensively validated in both simulation and real experiments.

It is time-consuming for acquiring complete data by fully phase encoding in two orthogonal directions along with one frequency encoding direction. Undersampling in the 3D k-space is promising in accelerating such 3D MRI process. Though 3D undersampling can be conducted according to predefined probability density, the density based method is not optimal. Because of the large amount of 3D data and computational cost, it is challenging to perform data-driven and learning-based 3D undersampling and subsequent 3D reconstruction. To tackle this challenge, this paper proposes a deep neural network called EEUR-Net, realized by optimizing specific undersampling patterns for the fully-sampled 3D k-space data. Innovatively, our undersampling algorithm employs an end-to-end deep learning approach to optimize phase encoding patterns and uses a 3D U-Net for image reconstruction of undersampled data. Through end-to-end training, we obtain an optimized 3D undersampling pattern, which significantly enhances the quality of the reconstructed image under the same acceleration factor. A series of experiments on a knee MRI dataset demonstrated that, in comparison to standard random uniform, radial, Poisson and equispaced Cartesian undersampling schemes, our end-to-end learned undersampling pattern considerably improves the reconstruction quality of undersampled MRI images.

The aeroelastic response of lightweight low-speed aircraft with slender wings under extreme flow turbulence intensity is not well-understood. Experiments on a commmercial unmanned aerial vehicle (UAV) with a 3-meter wingspan and aspect ratio of 13.6 were performed in a large open-return wind tunnel with extreme flow turbulence intensity of T∞≈10%. The structural dynamics of the wing in the bending mode and the flow beneath the wing to capture the effect of aileron deflection was measured using laser displacement sensors and tomographic particle image velocimetry (PIV) techniques. The unsteady lift produced by the wing was also measured using a high-capacity load cell at an angle of attack, α of 2 degrees for three freestream velocities U∞ of 13.4 m/s, 17.9 m/s, and 26.8 m/s representing the UAV’s stall through cruise speed. It was found that high flow turbulence intensity with large integral length scales relative to the wing chord plays a dominant role in the large unsteady lift and wing displacements measured. The power spectral density (PSD) of wing structural vibration shows that flow shedding from the wing and the integral length scales have a significant impact on the overall power inherent in the bending vibration of the wing. Computation of vorticity iso-surfaces in the flow measurement volume surrounding the aileron reveal a striking observation; aileron deflection, δa of 10∘becomes less effective in producing additional spanwise vorticity, which is proportional to circulation and lift, at U∞ of 26.8 m/s because the freestream already has elevated levels of vorticity. A paradigm shift in design is suggested for light aircraft structures with slender wings operating in highly turbulent flow, which is to employ multiple control surfaces in order to respond to this flow and mitigate large bending or torsion displacements and the probability of structural failure.

Chest X-ray (CXR) is the first tool globally employed to detect cardiopulmonary pathologies. These acquisitions are highly affected by scattered photons due to the large field-of-view required. Scatter in CXRs introduces background in the images, which reduces their contrast. We developed three deep learning-based models to estimate and correct the scatter contribution to CXRs. We used a Monte Carlo (MC) ray-tracing model to simulate CXRs from human models obtained from CT scans using different configurations (depending on the availability of dual-energy acquisitions). The simulated CXRs contained the separated contribution of direct and scattered X-rays in the detector. These simulated datasets were then used as the reference for the supervised training of several NN. Three NN models (single and dual energy) were trained with the MultiResUNet architecture. The performance of the NN models was evaluated on CXRs obtained with the MC code from chest CTs of patients affected by COVID-19. The results showed that the NN models were able to estimate and correct the scatter contribution to CXRs with an error &lt;5%, being robust to variations in the simulation setup and improving the contrast in soft tissue. The single-energy model was tested with real CXRs, providing robust estimations of the scatter-corrected CXRs.

Employing low tensor rank decompositions in image inpainting has attracted increasing attention. This paper exploits a novel tensor-augmentation schemes to transform an image (a low-order tensor) to a higher-order tensor without changing the total number of pixels. The developed augmentation schemes enhance the low-rankness of an image under three tensor decompositions: matrix SVD, tensor train (TT) decomposition, and tensor singular value decomposition (t-SVD). By exploiting the schemes, we solve the image inpainting problem with three low-rank con-strained models which use the matrix rank, TT rank, and tubal rank as constrained priors re-spectively. The tensor tubal rank and tensor train multi-rank are developed from t-SVD and TT decomposition respectively. We exploit efficient ADMM algorithms for solving the three models. Experimental results demonstrate that our methods are effective for image inpainting and supe-rior to numerous close methods.

Underwater image processing faces significant challenges due to the absorption and scattering of light as it travels through water. A self-supervised learning network based on the self-attention mechanism is proposed for underwater visual applications in this paper. The goal is to improve the effectiveness and stability of feature extraction from underwater images. By incorporating self-attention mechanism, the sensitivity of the original network architecture to degraded features of blurred underwater images is enhanced. The network proposed in this paper is trained using transfer learning and evaluated on various underwater image datasets. With distributive, quantitative, and qualitative advantages compared with other methods, experimental results have demonstrated that the algorithm presented shows a slighter decline in feature extraction ability for blurred images as the turbidity of water increases.

(1) Background: The imaging energy range of a typical Compton camera is limited due to the fact that scattered gamma photons are seldom fully absorbed when the incident energies are above 3 MeV. Further improving the upper energy limit of gamma-ray imaging has important application significance in active interrogation of special nuclear materials and chemical warfare agents, as well as range verification of proton therapy; (2) Methods: To realize gamma-ray imag-ing in a wide energy range of 0.3~7 MeV, a principle prototype, named a portable three-layer Compton camera, was developed using the scintillation detector that consist of an silicon photo-multiplier array coupled with a Gd3Al2Ga3O12:Ce pixelated scintillator array. Implemented in a list-mode maximum likelihood expectation maximization algorithm, a far-field energy-domain imaging method based on two-interaction and three-interaction events was applied to estimate the initial energy and spatial distribution of gamma-ray sources. The simulation model of detec-tors was established based on Monte Carlo simulation toolkit Geant4. The reconstructed images of a 133Ba, a 137Cs and a 60Co point-like sources were successfully obtained with our prototype in laboratory tests and compared with simulation studies; (3) Results: The proportion of effective imaging events accounted for about 2%, which made our prototype realize reconstructing the distribution of a 0.05 μSv/h 137Cs source in 10 seconds. The angular resolution for resolving two 137Cs point-like sources was 15°. Additional simulated imaging of the 6.13 MeV gamma-rays from 14.1 MeV neutron scattering with water preliminarily demonstrated the imaging capability for high incident energy; (4) Conclusions: We concluded that the prototype had good imaging per-formance in a wide energy range (0.3~7 MeV), which showed potential in several MeV gam-ma-ray imaging applications.

Breast cancer is a common malignant tumour and studies have shown that early and accurate detection is crucial for patients. With the maturity of medical imaging and deep learning development, significant progress has been made in breast cancer classification, which greatly improves the accuracy and efficiency of classification. This review focuses on deep learning, migration learning, GAN, and lifelong learning to elaborate and summarise the important roles arising from breast cancer detection. This review also examines the dataset and labeling issues required for breast cancer classification. In conclusion, at the end of the article, we look at future directions for breast cancer classification research, including cross-migration learning, multimodal data fusion, model interpretability, and lifelong learning, and also explore how to provide personalized treatment plans for patients.

Despite the impressive performance of deep neural networks on many different vision tasks, they have been known to be vulnerable to intentionally added noise to input images. To combat these adversarial examples (AEs), improving the adversarial robustness of models has emerged as an important research topic, and research has been conducted in various directions including adversarial training, image denoising, and adversarial purification. Among them, this paper focuses on adversarial purification, which is a kind of pre-processing that removes noise before AEs enter a classification model. The advantage of adversarial purification is that it can improve robustness without affecting the model’s nature, while another defense techniques like adversarial training suffer from a decrease in model accuracy. Our proposed purification framework utilizes a Convolutional Autoencoder as a base model to capture the features of images and their spatial structure. We further aim to improve the adversarial robustness of our purification model by distilling the knowledge from teacher models. To this end, we train two Convolutional Autoencoders (teachers), one with adversarial training and the other with normal training. Then, through ensemble knowledge distillation, we transfer the ability of denoising and restoring of original images to the student model (purification model). Our extensive experiments confirm that our student model achieves high purification performance(i.e., how accurately a pre-trained classification model classifies purified images). The ablation study confirms the positive effect of our idea of ensemble knowledge distillation from two teachers on performance.

Digital Image Correlation (DIC) is an optical technique used to measure surface displacements and strains in materials and structures. This technique has demonstrated significant utility in structural examination and monitoring. This manuscript offers a comprehensive review of contemporary research and applications that have leveraged the DIC technique in laboratory-based structural tests. The reviewed works encompass a broad spectrum of structural components, such as concrete beams, columns, pillars, masonry walls, infills, composite materials, structural joints, steel beams, slabs, and other structural elements. These investigations have underscored the efficacy of DIC as a metrological instrument for precise quantification of surface deformation and strain in these structural components. Moreover, the constraints of the DIC technique have been highlighted, especially in scenarios involving extensive or complex test configurations. Notwithstanding these constraints, the DIC methodology has validated its effectiveness as a strain measurement instrument, offering numerous benefits such as non-invasive operation, full-field measurement capability, high precision, real-time surveillance, and compatibility with integration into other measurement instruments and methodologies.

Eight lesions were analyzed using some algorithms of Intelligence Artificial: basal cell carcinoma (BCC), squamous cell carcinoma (SCC), melanoma (MEL), actinic keratosis (AK), benign keratosis (BKL), dermatofibromas (DF), melanocytic nevi (NV), and vascular lesions (VASC). This manuscript presents the possibility of using concatenated signatures (instead of images) obtained from different integral transforms, such as Fourier, Mellin, and Hilbert, to classify skin lesions. Eleven other Artificial Intelligence models were applied so that eight skin lesions could be classified by analyzing the particular signatures of each lesion. The database was randomly divided into 80%–20% for the training and test datasets images, respectively. The metrics that are being reported are accuracy, sensitivity, specificity, and precision. Each case was repeated 30 times to avoid bias, according to the central limit theorem in this work, and the average and ±standard deviation were reported. Although all the results were very satisfactory, the best average mark for the eight lesions analyzed was obtained using the Subspace KNN model, where the metrics for the test were 99.98% accuracy, 99.96% sensitivity, 99.99% specificity, and 99.95% precision.

Star image registration is the most important step in the application of astronomical image differencing, stacking and mosaicking, which requires high robustness, accuracy and real--time of the algorithm, but there is no high--performance registration algorithm in this field. In this paper, we propose a star image registration algorithm that relies only on radial module features (RMF) and rotation angle features (RAF), which has excellent robustness, high accuracy, and good real--time performance. The test results on a large amount of simulated and real data show that the comprehensive performance of the proposed algorithm is significantly better than the four classical baseline algorithms in the presence of rotation, insufficient overlapping area, false stars, position deviation, magnitude deviation and complex sky background, which is a more ideal star image registration algorithm.

Synthetic Aperture Radar (SAR) is an active sensor that uses microwave for sense, it is unrestricted by weather and illumination conditions, and it can observe targets all day and weather. Aircraft targets are important monitoring objects in military and civilian fields, and how to efficiently detect and recognize aircraft targets is an important topic in the field of SAR image interpretation. Based on the features of SAR images, such as complex background, high resolution, and multi-scale, we proposed an improved method based on YOLOv5s. Firstly, this paper proposed the structure of the multi-scale receptive field and channel attention fusion, which is applied to the shallow layer of the backbone of YOLOv5s, it can adjust the weights of the multi-scale receptive field during the training process to enhance the extraction ability of feature information. Secondly, we proposed four de-coupled detection heads to replace the original part in YOLOv5s, which can improve the efficiency and accuracy of SAR image interpretation for small targets. Thirdly, in the case of the limited amount of SAR images, this paper proposed multi methods of data augmentation, which can enhance the diversity and generalization of the network. Fourthly, this paper proposed the K-means++ to re-place the original K-means to improve the network convergence speed and detection accuracy. Finally, Experiments demonstrate that the improved YOLOv5s can enhance the accuracy of SAR image interpretation by 9.3%, and the accuracy of small targets is improved more obviously, reaching 13.1%.

To compare the quality of low-energy VMIs obtained with three DECT platforms according to phantom diameter. Three sections of the Mercury Phantom 4.0 were scanned on two generations of split-filter CTs (SFCT-1st and SFCT-2nd) and on one Dual-source CT (DSCT). Noise power spectrum (NPS), task-based transfer function (TTF) and detectability index (d’) were assessed on VMIs from 40 to 70keV. Highest noise magnitude values were found with SFCT-1st and noise magnitude was higher with DSCT than with SFCT-2nd for 26-cm (10.2±1.3%) and 31-cm (7.0±2.5%), and the opposite for 36-cm (-4.2±2.5%). The highest average NPS spatial frequencies and TTF values at 50% (f50) values were found with DSCT. For all energy levels, f50 values were higher with SFCT-2nd than SFCT-1st for 26-cm (3.2±0.4%) and the opposite for 31-cm (-6.9±0.5%) and 36-cm (-5.6±0.7%). The lowest d’ values were found with SFCT-1st. For all energy levels, d’ values were lower with SFCT-2nd than with DSCT for 26-cm (-6.2±0.7%), similar for 31-cm (-0.3±1.9%) and higher for 36-cm (5.4±2.7%). In conclusion, compared to SFCT-1st, SFCT-2nd exhibited lower noise magnitude and higher detectability values. Compared with DSCT, SFCT-2nd had lower noise magnitude and higher detectability for the 26-cm but the opposite was true for the 36-cm.

This paper proposes an approach for image data extraction and driving behavior analysis using geographic information and driving data. Information derived from geographic and global positioning systems was used for image data extraction. In addition, we used an onboard diagnostic II and a controller area network bus logger to record driving data for driving behavior analysis. Driving behavior was analyzed using sparse automatic encoders and data exploration to detect abnormal and aggressive behavior. A regression analysis was used to derive the relationship between aggressive driving behavior and road facilities. Several traffic improvements were proposed for specific intersections and roads. The results indicate that lane ratios without lane markings and with straight lane markings are important features that affect aggressive driving behaviors.

Deep learning has been introduced to single-image super-resolution (SISR) in the last decade. These techniques have taken over the benchmarks of SISR tasks. Nevertheless, most architectural designs necessitate substantial computational resources, leading to a prolonged inference time on embedded systems or rendering them infeasible for deployment. This paper presents a comprehensive survey of plausible solutions and optimization methods to address this problem. Then, we propose a pipeline that aggregates the latter in order to enhance the inference time without significantly compromising the perceptual quality. We investigate the effectiveness of the proposed method on a lightweight GAN-based perceptual-oriented model as a case study. The experimental results show that our proposed method leads to significant improvement in the inference time on both Desktop and Jetson Xavier NX, especially for higher resolution input sizes on the latter, thereby making it deployable in practice.

In the context of Shared Autonomous Vehicles, the need to monitor the environment inside the car will be crucial. This article focuses on the application of deep learning algorithms to detect objects, namely lost/forgotten items to inform the passengers, and aggressive items to monitoring if violent actions may arise between passengers. For object detection algorithms was used public datasets (COCO and TAO) to train state-of-the-art algorithms, such as YOLOv5. For violent action detection was used the MoLa InCar dataset to train on state-of-the-art algorithms such as I3D, R(2+1)D, SlowFast, TSN and TSM. At the end an embedded automotive solution was used to demonstrate both methods running in real-time.

This review focuses on developments that quantify grain quality with a range of spectral sensors in an on-farm setting. If the application of sensor technologies were expanded and adopted on-farm, growers could identify the impact and manage the harvesting operation to meet a range of quality targets and provide an economic advantage to the farming enterprise.

Owing to the advantages of cost-effectiveness, compactness, and the avoidance of complicated camera synchronization, single-camera three-dimensional (3D) digital image correlation (DIC) techniques have gained increasing attention for deformation measurement of materials and structures. In the traditional single-camera 3D-DIC system, the left and right view images can be recorded by a single camera using a diffraction grating, a bi-prism or a set of planar mirrors. To further improve the measurement accuracy of single-camera 3D-DIC, this paper introduces a single-camera four-view imaging technique by installing a pyramidal prism in front of the camera. The 3D reconstruction of the measured points before and after deformation is realized with eight governing equations induced by four views, and the strong geometric constraints of four views can help to improve the measurement accuracy. A static experiment, a rigid body translation experiment and a four-point bending experiment show that the proposed single-camera 3D-DIC method can achieve higher measurement accuracy than the dual-view single-camera 3D-DIC techniques and has advantages in reducing both random error and systematic error.

There is great interest in automatically detecting road weather and understanding its impacts on the overall safety of the transport network. This can, for example, support road condition-based maintenance or even serve as detection systems that assist safe driving during adverse climate conditions. In computer vision, previous work has demonstrated the effectiveness of deep learning in predicting weather conditions from outdoor images. However, training deep learning models to accurately predict weather conditions using real-world road-facing images is difficult due to: (1) the simultaneous occurrence of multiple weather conditions; (2) imbalanced occurrence of weather conditions throughout the year; and (3) road idiosyncrasies, such as road layouts, illumination, road objects etc. In this paper, we explore the use of focal loss function to force the learning process to focus on weather instances that are hard to learn with the objective to help address data imbalance. In addition, we explore the attention mechanism for pixel based dynamic weight adjustment to handle road idiosyncrasies using state-of-the-art vision transformer models. Experiments with a novel multi-label road weather dataset show that focal loss significantly increases the accuracy of computer vision approaches for imbalanced weather conditions. Furthermore, vision transformers outperforms current state-of-the-art convolutional neural networks in predicting weather conditions with a validation accuracy of 92% and F1-score of 81.22%, which is impressive considering the imbalanced nature of the dataset.

Fog, haze, or smoke are usual atmospheric phenomena that dramatically compromise the overall visibility of any scene, critically affecting features such as illumination, contrast, and contour detection of objects. The decrease in visibility compromises the performance of computer vision algorithms such as pattern recognition and segmentation, some of them very relevant for decision-making for the security or autonomous vehicle industries. Several dehazing methods have been proposed, however, to the best of our knowledge, all proposed metrics compare the defogged image to its ground truth for evaluation of the defogging algorithms, or need to estimate parameters through physical models. This fact hinders progress in the field as obtaining proper ground truth images is costly and time-consuming, and physical parameters greatly depend on the scene conditions. This paper aims to tackle this issue by proposing a contour-based metric for image defogging evaluation that does not need a ground truth image. The proposed metric only requires the original hazy RGB image and the RGB image after the defogging procedure. A comparison of the proposed metric with metrics currently used in the NTIRE 2018 defogging challenge is performed to prove its effectiveness in a general situation, showing comparable results to conventional metrics.

Over recent years, the importance of patent literature has become more recognized in the academic setting. In the context of artificial intelligence, deep learning, and data sciences, patents are relevant to not only industry but also academe and other communities. In this article, we focus on deep tomographic imaging and perform a preliminary landscape analysis of the related patent literature. Our search tool is PatSeer . Our patent bibliometric data is summarized in various figures and tables. In particular, we qualitatively analyze key deep tomographic patent literature.

Functional Data Analysis (FDA) is a relatively new field of statistics dealing with data expressed in the form of functions. FDA methodologies can be easily extended to the study of imaging data, an application proposed in Wang et al. (2020), where the authors settle the mathematical groundwork and properties of the proposed estimators. This methodology allows for the estimation of mean functions and simultaneous confidence corridors (SCC), also known as simultaneous confidence bands, for imaging data and for the difference between two groups of images. This is especially relevant for the field of medical imaging, as one of the most extended research setups consists on the comparison between two groups of images, a pathological set against a control set. FDA applied to medical imaging presents at least two advantages compared to previous methodologies: it avoids loss of information in complex data structures and avoids the multiple comparison problem arising from traditional pixel-to-pixel comparisons. Nonetheless, computing times for this technique have only been explored in reduced and simulated setups (Arias-López et al., 2021). In the present article, we apply this procedure to a practical case with data extracted from open neuroimaging databases and then measure computing times for the construction of Delaunay triangulations, and for the computation of mean function and SCC for one-group and two-group approaches. The results suggest that previous researcher has been too conservative in its parameter selection and that computing times for this methodology are reasonable, confirming that this method should be further studied and applied to the field of medical imaging.

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.

Obtaining classification confidence at the pixel level is a challenging task for accuracy assessment in remote sensing image classification. Among the various methods for estimating classification confidence at the pixel level, interpolation-based methods have drawn special attention in the literature. Even though they have been widely recognized in the literature, their usefulness has not been rigorously evaluated. This paper conducts a comprehensive evaluation of three interpolation-based methods: local error matrix method, bootstrap method, and geostatistical method. We applied each of the three methods to three representative datasets with different spatial resolutions, spectral bands, and the number of classes. We then derive the estimated classification confidence and true classification confidence and compared the results with each other using both exploratory data analysis (bi-histogram) and statistical analysis (Willmott's d and Binned classification quality). The results indicate that the three interpolation methods provide some interesting insights on various aspects of estimating per-pixel classification confidence. Unfortunately, the interpolation assumes that classification confidence is smooth across the space, which is usually not true in practice. In other words, interpolation-based methods have limited practical use.

An evaluation of image contrast in a controlled foggy environment has been made by comparison between intensity imaging and different modes of polarimetric imaging. A small-scale fog chamber has been designed and constructed to create the necessary controlled foggy environment. A linear polarimetric camera of division of focal plane and a linearly polarized light source has been used for performing the experiments with polarized light. In order to evaluate the image contrast of the different imaging modes, the Michelson’s contrast of samples of different materials relative to their background has been calculated. The higher the image contrast, the easier is to detect and segment the targets of interest that are surrounded by fog. It has been proved quantitatively that in the studied situations polarimetric images present an improvement in contrast when compared to conventional intensity images.

The analysis of microscopy images has always been an important yet time consuming process in in materials science. Convolutional Neural Networks (CNNs) have been very successfully used for a number of tasks, such as image segmentation. However, training a CNN requires a large amount of hand annotated data, which can be a problem for material science data. We present a procedure to generate synthetic data based on ad-hoc parametric data modelling for enhancing generalization of trained neural network models. Especially for situations where it is not possible to gather a lot of data, such an approach is beneficial and may enable to train a neural network reasonably. Furthermore, we show that targeted data generation by adaptively sampling the parameter space of the generative models gives superior results compared to generating random data points.

Satellite Image Time Series (SITS) is a data set which includes satellite images across several years with a high acquisition rate. Radiometric normalization is a fundamental and important preprocessing method for remote sensing applications using SITS due to the radiometric distortion caused by noise between images. Normalizing the subject image based on the reference image is a general strategy when using traditional radiometric normalization methods to normalize multi-temporal imagery (usually two scenes or three scenes in different time phases). However, these methods are not suitable for calibrating SITS because they cannot minimize the radiometric distortion between any pair of images in SITS. The existing relative radiometric normalization methods for SITS are based on linear assumptions which cannot effectively reduce nonlinear radiometric distortion caused by continuously changing noise in SITS. To overcome this problem and obtain a more accurate SITS, this study proposed a Nonlinear Radiometric Normalization Model (named NMAG) for SITS based on Artificial Neural Networks (ANN) and Greedy Algroithm (GA) . In this method, GA was used to determine the correction order of SITS and calculate the error between the image to be corrected and normalized images, which avoided the selection of a single reference image. ANN was used to obtain the optimal solution of error function, which minimized the radiometric distortion between different images in SITS. SITS composed of 21 Landsat-8 images in Tianjin City from October 2017 to January 2019 were selected to test the method. We compared NMAG with other two contrasting method (refered as CM1 and CM2), and found that the average of root mean square error $(\mu_{RMSE})$ of NMAG $(497.22)$ was significantly smaller than those of CM1 $(641.39)$ and CM2 $(543.47)$, and the accuracy of normalized SITS obtained by using NMAG has increased by 22.4\% and 8.5\% toward CM1 and CM2, respectively. These experimental results confirmed the effectiveness of NMAG in the reduction of radiometric distortion caused by continuously changing noise between images in SITS.

CNN-based methods have dominated the field of aerial scene classification for the past few years. While achieving remarkable success, CNN-based methods suffer from excessive parameters and notoriously rely on large amounts of training data. In this work, we introduce few-shot learning to the aerial scene classification problem. Few-shot learning aims to learn a model on base-set that can quickly adapt to unseen categories in novel-set, using only a few labeled samples. To this end, we proposed a meta-learning method for few-shot classification of aerial scene images. First, we train a feature extractor on all base categories to learn a representation of inputs. Then in the meta-training stage, the classifier is optimized in the metric space by cosine distance with a learnable scale parameter. At last, in the meta-testing stage, the query sample in the unseen category is predicted by the adapted classifier given a few support samples. We conduct extensive experiments on two challenging datasets: NWPU-RESISC45 and RSD46-WHU. The experimental results show that our method yields state-of-the-art performance. Furthermore, several ablation experiments are conducted to investigate the effects of dataset scale, the impact of different metrics and the number of support shots; the experiment results confirm that our model is specifically effective in few-shot settings.

For our research on droplet deformation and breakup in scaled high-pressure homogenizing units we developed a pressure stable inline droplet generator. It consists of an optically accessible flow channel with a combination of stainless steel and glass capillaries and a 3D printed orifice. The droplet size is determined online by live image analysis. The influence of the orifice diameter, the mass flow of the continuous phase and the mass flow of the disperse phase on the droplet diameter was investigated. Furthermore, the droplet detachment mechanisms were identified. Droplet diameters with small diameter fluctuation between 175 µm and 500 µm could be realized, which allows a precise adjustment of the Ca and We Number in the subsequent scaled high pressure homogenizer disruption unit. The determined influence of geometry and process parameters on the resulting droplet size and droplet detachment mechanism agreed well with literature on microfluidics. Furthermore, droplet trajectories in an exemplary scaled high-pressure homogenizer disruption unit are presented which show that the droplets can be reinjected on a trajectory close to the center axis or close to the wall, which should result in different stresses on the droplets.

Image registration has an imperative role in medical imaging. In this work, a grey-wolf optimizer (GWO) based non-rigid demons registration is proposed to support the retinal image registration process. A comparative study of the proposed GWO-based demons registration framework with cuckoo search, firefly algorithm, and particle swarm optimization- based demons registration is conducted. In addition, a comparative analysis of different demons registration methods, such as Wang’s demons, Tang’s demons, and Thirion’s demons which are optimized using the proposed GWO is carried out. The results established the superiority of the GWO-based framework which achieved 0.9977 correlation, and fast processing compared to the use of the other optimization algorithms. Moreover, GWO-based Wang’s demons performed better accuracy compared to the Tang’s demons and Thirion’s demons framework. It also achieved the best less registration error of 8.36×10^-5.

Terrestrial laser scanning (TLS) provides detailed three-dimensional representation of the surrounding forest structure. However, due to close-range hemispherical scanning geometry, the ability of TLS technique to comprehensively characterize all trees and especially the upper parts of forest canopy is often limited. In this study, we investigated how much forest characterization capacity can be improved in managed Scots pine (Pinus sylvestris L.) stands if TLS point cloud is complemented with a photogrammetric point cloud acquired from above the canopy using unmanned aerial vehicle (UAV). In this multisensorial (TLS+UAV) close-range sensing approach, the used UAV point cloud data was considered feasible especially in characterizing the vertical forest structure and improvements were obtained in estimation accuracy of tree height as well as plot-level basal-area weighted mean height (H_g) and mean stem volume (V_mean). Most notably the root mean square error (RMSE) in H_g improved from 0.88 m to 0.58 m and the bias improved from -0.75 m to -0.45 m with the multisensorial close-range sensing approach. However, in managed Scots pine stands the mere TLS captured also the upper parts of the forest canopy rather well. Both approaches were capable of deriving stem number, basal area, V_mean, H_g and basal area-weighted mean diameter with a relative RMSE less than 5.5% for all of the sample plots. Although the multisensorial close-range sensing approach mainly enhanced characterization of forest vertical structure in single-species, single-layer forest conditions, representation of more complex forest structures may benefit more from point clouds collected with sensors of different measurement geometries.

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.

A fast calculation method for the full parallax high-resolution hologram is proposed based on the elemental light field image (EI) rendering. A 3D object located near the holographic plane is firstly rendered as multiple EIs with a pinhole array. Each EI is interpolated and multiplied by a divergent sphere wave and interfered with a reference wave to form a hogel. Parallel acceleration is used to calculate the high-resolution hologram because calculation of each hogel is independent. A high-resolution hologram with the resolution of 20,0000×20,0000 pixels is calculated only within 8 minutes. Full parallax high-resolution 3D displays are realized by optical reconstructions.

Alnus glutinosa is important woody plant in Lithuanian forest ecosystems. Knowledge of fluorescence properties of black alder pollen is necessary for scientific and practical purposes. By the results of the study we aimed to evaluate possibilities of identifying Alnus glutinosa pollen fluorescence properties by modeling ozone effect and applying two different fluorescence-based devices. To implement experiments, black alder pollen was collected in a typical habitat during the annual flowering period in 2018-2019. There were three groups of experimental variants, which differed in the duration of exposure to ozone, conditions of pollen storage before the start of the experiment, and the experiment start time. Data for pollen fluorescence analysis were collected using two methods. The microscopy method was used in order to evaluate the possibility of employing image analysis systems for investigation of pollen fluorescence. The second data collection method is related to the automatic device identifying pollen in real-time, which uses the fluorescence method in the pollen recognition process. Data were assessed employing image analysis and principal component analysis (PCA) methods. Digital images of ozone-exposed pollen observed under the fluorescence microscope showed the change of the dominant green colour towards the blue spectrum. Meanwhile, the automatic detector detects more pollen whose fluorescence is at the blue light spectrum. It must be noted that assessing pollen fluorescence several months after exposure to ozone, no effect of ozone on fluorescence remains.

The study concentrates on researching possibilities of using computer image analysis and neural modeling in order to assess selected quality discriminants of spray-dried chokeberry powder. The aim of the paper is quality identification of chokeberry powders on account of their highest dying power, the highest bioactivity as well as technologically satisfying looseness of powder. The article presents neural models with vision technique backed up by devices such as digital camera as well as electron microscope. Reduction in size of input variables with PCA has influence on improving the processes of learning data sets, thus increasing effectiveness of identifying chokeberry fruit powders included in digital pictures, which is shown in the results of the conducted research. The effectiveness of image recognition are presented by classifying abilities as well as low Root Mean Square Error (RMSE), for which the best results are achieved with typology of network type Multi-Layer Perceptron (MLP). The selected networks type MLP are characterized by the highest degree of classification at 0.99 and RMSE at 0.11 at most at the same time.

This paper discusses the metonymic uses of the greek deictic adverbs εδώ [here] and εκεί [there] in the language of politics. The paper draws examples from political speeches which taken place in the Hellenic Parliament during 2011 and discussed the financial situation of Greece during that time. The paper discusses the multiple senses of these deictic adverbs and suggests that the temporal and spatial denotations of εδώ and εκεί are subject to image schemas. It is argued that the image schemas in which εδώ and εκεί are rooted have a metonymic basis. The paper also suggests that the spatio-temporal senses of εδώ and εκεί go beyond their deictic function due to their metonymic basis.

Spatio-temporal Land-Use and Land-Cover (LULC) changes have been affecting geo-environmental and climate change globally. This study aims to analyze LULC changes in Bahir Dar city and its surrounds. Landsat 5 TM (1987), Landsat 7 ETM+ (2002) and Landsat 8 OLI (2017) and SPOT images, and aerial photographs, master plan map and Google Earth Landsat images were used to analyze changes. In Bahir Dar city and its surrounds, LULC has been changing in space and time. During 1987-2017, more than 50% of the study area was covered with cropland. Settlement areas have increased from 3.3% in 1987 to 9.13% in 2017. However, wetland vegetation, shrubland, grassland, forest, and waterbodies have degraded. These changes are mainly attributed to population growth and its effect on the environment. Land-use and land-cover is a serious problem and it causes land and environmental degradation, climate change and loss of the biological environment.

The identification of bloodstain is one of the most important approaches in obtaining evidence in criminalistics. A threshold method based on spectral co-efficient and interclass variance is proposed in this paper, it is a non-contact, non-destructive method for quickly identifying bloodstains. The spectra of bloodstains and other suspected substances were all extracted from their hyper-spectral image. Then calculate the correlation coefficients of these spectral and interclass variances, analyze the differences between substances. The best blood recognition threshold was determined as 0.9. After preprocessing for eliminating systematic errors, experiments with the threshold 0.9 are carried out to identify bloodstains on the calico and red T-shirt. The method can remarkably identify the bloodstain from other non-blood substances both quickly and efficiently. The blood extraction rate can reach to 93.35% and 89.19%, respectively. It is an important step toward the implementation of bloodstain non-contact and non-destructive identification in forensic casework.

During the last 25 years the scientific community has coexisted with the most fascinating protocol due to Quantum Physics: quantum teleportation (QTele), which would have been impossible if quantum entanglement, so questioned by Einstein, did not exist. In this work, a complete architecture for the teleportation of Computational Basis States (CBS) is presented. Such CBS will represent each of the possible 24 classical bits commonly used to encode every pixel of a 3-color-channel-image (red-green-blue, or cyan-yellow-magenta). For this purpose, a couple of interfaces: classical-to-quantum (Cl2Qu) and quantum-to-classical (Qu2Cl) are presented with two versions of the teleportation protocol: standard and simplified.

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.

Detection of the region of interest (ROI) is a critical step in laryngeal image analysis for the delineation of glottis contour. The process can improve both computational efficiency and accuracy of the image segmentation task, which will facilitate subsequent analysis and characterization of the vocal fold vibration as it correlates with voice quality and pathology. This study aims to develop machine learning based approaches for automatic detection of ROI for glottis image sequences captured by high-speed video-endoscopy (HSV), a clinical laryngeal imaging modality. In particular, we first applied the supporting vector machine (SVM) method using histogram of oriented gradients (HOG) feature descriptor, and second, trained a convolutional neural network (CNN) model for this task. Comparisons are made for both approaches in terms of accuracy of recognition and computation time.

In addition to image filtering in the spatial and frequency domains, fractal characteristics induced algorithms offers considerable flexibility in the design and implementations of image processing solutions in areas such as image enhancement, image restoration, image data compression and spectrum of applications of practical interests. Facing up to a real-world problem of identifying workpiece surface defects, a generic adaptive fractal filtering algorithm is proposed, which shows advantages on the problems of target recognition, feature extraction and image denoising at multiple scales. First, we reveal the physical principles underlying between signal SNR and its representative fractal dimension indicative parameters, validating that the fractal dimension can be used to adaptively obtain the image features. Second, an adaptive fractal filtering algorithm (Abbreviated as AFFA) is proposed according to the identified correlation between the image fractal dimensions and the scales of objects, and it is verified by a benchmarking image processing case study. Third, by using the proposed fractal filtering algorithm, surface defects on a flange workpiece are identified. Compared to conventional image processing algorithms, the proposed algorithm shows superior computing simplicity and better performance Numerical analysis and engineering case studies show that the fractal dimension is eligible for deriving an adaptive filtering algorithm for diverse-scale object identification, and the proposed AFFA is feasible for general application in workpiece surface defect detection.

Targets detection in synthetic aperture radar (SAR) remote sensing images, which is a fundamental but challenging problem in the field of satellite image analysis, plays an important role for a wide range of applications and is receiving significant attention in recent years. Besides, the ability of human visual system to detect visual saliency is extraordinarily fast and reliable. However, computational modeling of SAR image scene still remains a challenge. This paper analyzes the defects and shortcomings of traditional visual models applied to SAR images. Then a visual attention model designed for SAR images is proposed. The model draws the basic framework of classical ITTI model; selects and extracts the texture features and other features that can describe the SAR image better. We proposes a new algorithm for computing the local texture saliency of the input image, then the model constructs the corresponding saliency maps of features; Next, a new mechanism of feature fusion is adopted to replace the linear additive mechanism of classical models to obtain the overall saliency map; Finally, the gray-scale characteristics of focus of attention (FOA) in saliency map of all features are taken into account, our model choose the best saliency representation, Through the multi-scale competition strategy, the filter and threshold segmentation of the saliency maps can be used to select the salient regions accurately, thereby completing this operation for the visual saliency detection in SAR images. In the paper, several types of satellite image data, such as TerraSAR-X (TS-X), Radarsat-2, are used to evaluate the performance of visual models. The results show that our model provides superior performance compared with classical visual models. By further contrasting with the classical visual models, Our model reduce the false alarm caused by speckle noise, and its detection speed is greatly improved, and it is increased by 25% to 45%.

The use of unmanned aerial vehicle (UAV) during emergency response of a disaster has been widespread in recent years and the terrain images captured by the cameras on board these vehicles are significant sources of information for such disaster monitoring operations. Thus, analyzing such images are important for assessing the terrain of interest during such emergency response operations. Further, these UAVs are mainly used in disaster monitoring systems for the automated deployment of sensor nodes in real time. Therefore, deploying and localizing the wireless sensor nodes optimally, only in the regions of interest that are identified by segmenting the images captured by UAVs, hold paramount significance thereby effecting their performance. In this paper, the highly effective nature-inspired Plant Growth Simulation Algorithm (PGSA) has been applied for the segmentation of such terrestrial images and also for the localization of the deployed sensor nodes. The problem is formulated as a multi-dimensional optimization problem and PGSA has been used to solve it. Furthermore, the proposed method has been compared to other existing evolutionary methods and simulation results show that PGSA gives better performance with respect to both speed and accuracy unlike other techniques in literature.

Precision agriculture is a farm management technology that involves sensing and then responding to the observed variability in the field. Remote sensing is one of the tools of precision agriculture. The emergence of small unmanned aerial vehicles (sUAV) have paved the way to accessible remote sensing tools for farmers. This paper describes the comparison of two popular off-the-shelf sUAVs: 3DR Iris and DJI Phantom 2. Both units are equipped with a camera gimbal attached with a GoPro camera. The comparison of the two sUAV involves a hovering test and a rectilinear motion test. In the hovering test, the sUAV was allowed to hover over a known object and images were taken every second for two minutes. The position of the object in the images was measured and this was used to assess the stability of the sUAV while hovering. In the rectilinear test, the sUAV was allowed to follow a straight path and images of a lined track were acquired. The lines on the images were then measured on how accurate the sUAV followed the path. Results showed that both sUAV performed well in both the hovering test and the rectilinear motion test. This demonstrates that both sUAVs can be used for agricultural monitoring.

Historical maps constitute unique sources of retrospective geographic information. Recently, several map archives containing map series covering large spatial and temporal extents have been systematically scanned and made available to the public. The geographic information contained in such data archives allows extending geospatial analysis retrospectively beyond the era of digital cartography. However, given the large data volumes of such archives and the low graphical quality of older map sheets, the processes to extract geographic information need to be automated to the highest degree possible. In order to understand the salient characteristics, data quality variation, and potential challenges in large-scale information extraction tasks, preparatory analytical steps are required to efficiently assess spatio-temporal coverage, approximate map content, and spatial accuracy of such georeferenced map archives across different cartographic scales. Such preparatory steps are often neglected or ignored in the map processing literature but represent highly critical phases that lay the foundation for any subsequent computational analysis and recognition. In this contribution we demonstrate how such preparatory analyses can be conducted using classical analytical and cartographic techniques as well as visual-analytical data mining tools originating from machine learning and data science, exemplified for the United States Geological Survey topographic map and Sanborn fire insurance map archives.