General–purpose automatic speech recognition (ASR) systems have improved their quality and are being used for pronunciation assessment. However, the assessment of isolated short utterances, as words in minimal pairs for segmental approaches, remains an important challenge, even more for non-native speakers. In this work, we compare the performance of our own tailored ASR system (kASR) with the one of Google ASR (gASR) for the assessment of Spanish minimal pair words produced by 33 native Japanese speakers in a computer-assisted pronunciation training (CAPT) scenario. Participants of a pre/post-test training experiment spanning four weeks were split into three groups: experimental, in-classroom, and placebo. Experimental group used the CAPT tool described in the paper, which we specially designed for autonomous pronunciation training. Statistically significant improvement for experimental and in-classroom groups is revealed, and moderate correlation values between gASR and kASR results were obtained, beside strong correlations between the post-test scores of both ASR systems with the CAPT application scores found at the final stages of application use. These results suggest that both ASR alternatives are valid for assessing minimal pairs in CAPT tools, in the current configuration. Discussion on possible ways to improve our system and possibilities for future research are included.

We present improvements in n-best rescoring of code-switched speech achieved by n-gram augmentation as well as optimised pretraining of long short-term memory (LSTM) language models with larger corpora of out-of-domain monolingual text. In addition, we consider the application of large pretrained transformer-based architectures. Our experimental evaluation is performed on an under-resourced corpus of code-switched speech comprising four bilingual code-switched sub-corpora, each containing a Bantu language (isiZulu, isiXhosa, Sesotho, or Setswana) and English. We find in our experiments that, by combining n-gram augmentation with the optimised pretraining strategy, speech recognition errors are reduced for each individual bilingual pair by 3.51% absolute on average over the four corpora. Importantly, we find that even speech recognition at language boundaries improves by 1.14% even though the additional data is monolingual. Utilising the augmented n-grams for lattice generation, we then contrast these improvements with those achieved after fine-tuning pretrained transformer-based models such as distilled GPT-2 and M-BERT. We find that, even though these language models have not been trained on any of our target languages, they can improve speech recognition performance even in zero-shot settings. After fine-tuning on in-domain data, these large architectures offer further improvements, achieving a 4.45% absolute decrease in overall speech recognition errors and a 3.52% improvement over language boundaries. Finally, a combination of the optimised LSTM and fine-tuned BERT models achieves a further gain of 0.47% absolute on average for three of the four language pairs compared to M-BERT. We conclude that the careful optimisation of the pretraining strategy used for neural network language models can offer worthwhile improvements in speech recognition accuracy even at language switches, and that much larger state-of-the-art architectures such as GPT-2 and M-BERT promise even further gains.

This paper documents the development of a special case of multilingual Automatic Speech Recognition model, specifically tailored to attend two languages spoken by the majority of Latin America, Portuguese and Spanish. The bilingual model combines Language Identification and Speech Recognition developed with the Wav2Vec2.0 architecture and trained on several open and private speech datasets. In this model, the feature encoder is trained jointly for all tasks and different context encoders are trained for each task. The model is evaluated separately on two tasks: language identification and speech recognition. The results indicate that this model achieves good performance on speech recognition and average performance on language identification, training on a low quantity of speech material. The average accuracy of the language identification module on the MLS dataset is 66.75%. The average Word Error Rate in the same scenario is 13.89%, which is better than average 22.58% achieved by the commercial speech recognizer developed by Google.

Automatic Voice Query Service can extremely reduce the artificial cost, which could improve the response efficiency for users. The automatic speech recognition (ASR) is one of the important component in AVQS. However, many dialect areas in China make the AVQS have to response the multi-accented Mandarin users by single acoustic model in ASR. This problem severely limits the accuracy of ASR for multi-accented speech in the AVQS. In this paper, a new framework for AVQS is proposed to improve the accuracy of response. Firstly, the fusion feature including iVector and filterbank acoustic features is used to train the Transformer-CTC model. Secondly, the transformer-CTC model is used to construct the end-to-end ASR. Finally, key words matching algorithm for AVQS based on fuzzy mathematic theory is proposed to further improve the accuracy of response. The results show that the final accuracy in our proposed framework for AVQS arrives at 91.5%. The proposed framework for AVQS can satisfy the service requirement of different areas in mainland of China. This research has a great significance for exploring the application value of artificial intelligence in the real scene.

Over the past decades, overweight and obesity has become a global epidemic and the leading threat for death. To prevent the serious risk, an overweight or obese individual must apply a long-term weight-management strategy to control food intake and physical activities, which is however, not easy. Recently, with the advances of information technology, more and more people can use wearable devices and smartphones to obtain physical activity information, while they can also access various health-related information from Internet online social networks (OSNs). Nevertheless, there is a lack of an integrated approach that can combine these two methods in an efficient way. In this paper, we address this issue and propose a novel mobile-social framework for health recognition and recommendation, namely, H-Rec². The main ideas of H-Rec² include (1) to recognize the individual's health status using smartphone as a general platform, and (2) to recommend physical activity and food intake based on personal health information, life science principles, and health-related information obtained from OSNs. To demonstrate the potentials of the H-Rec² framework, we develop a prototype that consists of four important components: (1) an activity recognition module that senses physical activity using accelerometer, (2) a health status modeling module that applies a novel algorithm to generate personalized health status index, (3) a restaurant information collection module that collects relevant information from OSN, and (4) a restaurant recommendation module that provides personalized and context-aware recommendation. To evaluate the prototype, we conduct both objective and subjective experiments, which confirm the performance and effectiveness of the proposed system.

Satellite communication (Satcom) is an artificial geostationary satellite that facilitates a wide range of telecommunications. Considering its quality of service (QoS) and security is crucial in government/military applications. The most challenging situation for efficient Satcom is radio frequency interference (RFI) environment. Thus, it is necessary to ensure that transmissions are incorruptible or at least sense the quality of its spectrum. This paper presents a new method to recognize received signal characteristics using a hierarchical classification in a multi-layer perceptron neural network. We consider signal modulation and the type of RFI as the characteristics of a real-time video stream transmitted in the direct broadcast satellite. Four different modulation types are investigated in this study. Moreover, the combination of the communication signal with various kinds of interference and their effects on the classification method widely have been analyzed. Besides, two robust feature selection techniques have been developed to reduce the data-set dimensional, which leads to optimizing the classification process. The results show that the Genetic Algorithm (GA) slightly outperforms Principal Component Analysis (PCA) for feature selection. Furthermore, the robustness of the proposed techniques is assessed to detect unknown signals at different signal to noise ratios.

Recent studies have proven that synthetic aperture radar (SAR) automatic target recognition (ATR) models based on deep neural networks (DNN) are vulnerable to adversarial examples. However, existing attacks are easily failed in the case where adversarial perturbations cannot be fully fed to victim models. We call this situation perturbation offset. Moreover, since background clutter takes up most of the areas in SAR images and has low relevance to recognition results, fooling models with global perturbations is quite inefficient. This paper proposes a semi-whitebox attack network, called Universal Local Adversarial Network (ULAN), to generate universal adversarial perturbations (UAP) for the target regions of SAR images. In the proposed network, we calculate the model’s attention heatmaps through layer-wise relevance propagation (LRP), which is used to locate the target regions of SAR images that have high relevance to recognition results. In particular, we utilize a generator based on the U-Net to learn the mapping from noise to UAPs and craft adversarial examples by adding the generated local perturbations to target regions. Experiments indicate that the proposed method fundamentally prevents perturbation offset and achieves comparable attack performance to conventional global UAPs by perturbing only a quarter or less of SAR image areas.

This study evaluates whether a combination of photodiode sensor measurements, taken during laser powder bed fusion (L-PBF) builds, can be used to predict the resulting build quality via a purely data-based approach. We analyse the relationship between build density and features that are extracted from sensor data collected from three different photodiodes. The study uses a Singular Value Decomposition to extract lower-dimensional features from photodiode measurements, which are then fed into machine learning algorithms. Several unsupervised learning methods are then employed to classify low density (< 99% part density) and high density (≥ 99% part density) specimens. Subsequently, a supervised learning method (Gaussian Process regression) is used to directly predict build density. Using the unsupervised clustering approaches, applied to features extracted from both photodiode sensor data as well as observations relating to the energy transferred to the material, build density was predicted with up to 93.54% accuracy. With regard to the supervised regression approach, a Gaussian Process algorithm was capable of predicting the build density with a RMS error of 3.65%. The study shows, therefore, that there is potential for machine learning algorithms to predict indicators of L-PBF build quality from photodiode build-measurements. Moreover, the work herein describes approaches that are predominantly probabilistic, thus facilitating uncertainty quantification in machine-learnt predictions of L-PBF build quality.

Chronic venous disease (CVD) occurs in a substantial proportion of the world's population. If the onset of CVD looks like a cosmetic defect, then over time, it can develop into serious problems that require surgical intervention. The aim of the work is to use deep learning (DL) methods for automatic classification of the stage of CVD for self-diagnosis of a patient by using the image of the patient’s legs. The required for DL algorithms images of legs with CVD were obtained by using Internet Data Mining. For images preprocessing, the binary classification problem “legs - no legs” was solved based on Resnet50 with accuracy 0.998. The application of this filter made it possible to collect a data set of 11,118 good quality leg images with various stages of CVD. For classification of various stages of CVD according to CEAP classification, the multi classification problem was set and resolved by using two neural networks with completely different architecture - Resnet50 and DeiT. The model based on DeiT without any tuning shows better results than the model based on Resnet50 (precision = 0.770 (DeiT) and 0.615 (Resnet50)). To demonstrate the results of the work, a telegram bot was developed, in which fully functioning DL algorithms are implemented. This bot allows evaluating the condition of the patient's legs with a fairly good accuracy for the CVD classification.

Knowledge of the spatio-temporal occurrence of avalanche activity is critical for avalanche forecasting and hazard mapping. We present a near-real time automatic avalanche monitoring system that outputs detected avalanche polygons within roughly 10 min after Sentinel- 1 SAR data download. Our avalanche detection algorithm has an average probability of detection of 67.2 % with a false alarm rate averaging 45.9, with maximum POD's over 85 % and minimum FAR's of 24.9 % compared to manual detection of avalanches. The high variability in performance stems from the dynamic nature of snow in the Sentinel-1 data. After tuning parameters of the detection algorithm, we processed five years of Sentinel-1 images acquired over a 150 x 100 km large area in Northern Norway, with the best setup. Compared to a dataset of field-observed avalanches, 77.3 % were manually detectable. Using these manual detections as benchmark, the avalanche detection algorithm achieved an accuracy of 79 % with high POD's in cases of medium to large wet snow avalanches. For the first time, we can present a dataset of spatiotemporal avalanche activity over several winters from a large region. This unique dataset allows for research into the relationship between avalanche activity and triggering meteorological factors, mapping of avalanche prone areas and near-real time avalanche activity monitoring to assist public avalanche forecasting. Currently, the Norwegian Avalanche Warning Service is using our processing system for pre-operational use in three regions in Norway.

Multimodality requires integration of heterogeneous transportation data to construct a broad view of the transportation network. Many new transportation services are emerging with being isolated from previously existing networks. This lead them to publish their data sources to the web -- according to Linked Data Principles -- in order to gain visibility. Our interest is to use these data to construct an extended transportation network that links these new services to existing ones. The main problems we tackle in this article fall in the categories of automatic schema matching and data interlinking. We propose an approach that uses web services as mediators to help in automatically detect geospatial properties and map them between two different schemas. On the other hand, we propose a new interlinking approach that enables user to define rich semantic links between datasets in a flexible and customizable way.

The objective of this study was to evaluate biomarkers of heat stress (HS) from automatic milking system (AMS), the relationships between measurements of temperature-humidity index (THI) and reticulorumen ph and temperature and some automatic milking systems parameters in dairy cows (rumination time (RT), milk traits, body weight (BW) and consumption of concentrate (CC)) during summer period. The experiment was carried out on a dairy farm at 54.9587408, 23.784146. Lithuanian Black and White dairy cows (n=365) were selected. The cows were milked with Lely Astronaut® A3 milking robots with free traffic. The rations were calculated to meet physiological requirements of the animals. Daily milk yield, rumination time, body weight, milk fat and protein ratio were collected from the Lely T4C management program for analysis. The pH and temperature of the contents of cow reticulorumen were measured using specific smaX-tec boluses manufactured for animal care. The daily humidity and air temperature in the farm zone were obtained from the adjacent weather station (2 km away). According this study during HS, the higher THI had positive correlation with ML, which show tendencies to increase risk of mastitis, and decrease CC, RT, BW, MY, reticulorumen ph and F/P. Some biomarkers of HS can be milk yield, milk lactose, somatic cell count, concentrate conception, rumination time, body weight, reticulorumen ph and milk fat – protein ratio. We can recommend to monitoring these parameters in the herd management program to identify possibility of heat stress.

Automation of vegetable seedling transplanting has provided opportunities for saving labors and improving productivity. Some changes in seedling agronomy are necessary for efficient transplanting. In this study, the local nursery substrates were added with the herbaceous peat, the sphagnum peat and the coir peat, respectively. Effects of the new compound substrates were investigated on the seedling qualities and the root-substrate strength. In the results, we found that the addition of these horticultural mediums significantly affected the physiochemical properties of the original substrates. Under the same nursery conditions, some appropriate additions could promote the seedling growth. And the deficient or excessive additions were to inhibit the growing development of seedlings and their roots. The corresponding additions would also improve the structural characteristics of the root lumps. Generally, the nursery substrates added with the sphagnum peat were relatively optimized in contribution to the seedling qualities and the root-substrate strength. Especially as the commercial substrate and the sphagnum peat were mixed at the volume ratio of 2:1, the dry matter accumulation of seedlings was 2.18 times more than the original. Their root lumps had the best consolidation strength, which may be an effective application for the necessary qualities of seedlings for automatic transplanting.

As use of composite materials increases, the search for suitable automated processes gains relevance to guarantee production quality by ensuring uniformity of the process, minimizing the amount of generated scrap and reducing time and energy consumption. Limitations on production by traditional means such as hand lay-up, vacuum bagging and in-autoclave methods, tend not to be as efficient when the size and shape complexity of the part being produced increases, motivating the search for alternative processes such as the Automated Tape Laying (ATL). This work aims to describe the process of modelling and simulating a composite ATL with in situ consolidation by characterizing the machine elements, using the finite differences method in conjunction with energy balances, in order to create a digital twin of the process for further control design. The modelling approach implemented is able to follow the process dynamics when changes to the heating element are imposed as well as to predict the composite material temperature response, making it suitable to work as a digital twin of a production process using an ATL machine.

In this study pH, temperature of the contents of the forestomach of cows and cow activity were measured using specific smaXtec boluses manufactured for animal care. Rumination time, body weight, milk yield, milk fat/protein ratio, milk lactose, milk somatic cell count, milk electrical conductivity and conception of concentrates were registered with the help of Lely Astronaut® A3 milking robots. The following parameters were obtained: base excess in blood, partial carbon dioxide pressure, partial oxygen pressure, bicarbonate, hydrogen potential, total carbon dioxide carbon, base excess in extracellular fluid, sodium, calcium, potassium, packed cell volume, chlorides, hemoglobin concentration and lactate. According to the concentration of lactose in milk, cows were grouped into two groups: group 1 - milk lactose &lt;4.70% (n = 20), group 2 - milk lactose ≥ 4.70% (n = 15). Data of cows were also divided by milk fat and protein ratio: F/P&lt;1.2 (class 1), F/P=1.2 (class 2) and F/P&gt;1.2 (class 3). According our results we can conclude that inline registered milk lactose concentration can be used as indicator for the health status of fresh dairy cows. Cows with higher lactose concentration (≥ 4.70%) developed more activity (54.47%) and had less risk of mastitis (lover milk EC, and SCC) and metabolic disorders according to F/P. Cows with higher lactose concentration showed higher glucose concentrations. Low level of lactose can be used as indicator of mastitis (milk SCC ≥ 100 thousand/ml) and metabolic disorders according to F/P.

Auto checkout has received more and more attention in recent years and this system automatically generates a shopping bill by identifying the picture of the products purchased by the customers. However, the system is challenged by the domain adaptation problem, where each image of the training set contains only one commodity, whereas the test set is a collection of multiple commodities. The existing solution to this problem is to resynthesize the training images to enhance the training set. Then the composite images are rendered using CycleGAN to make the image distribution of the training set and the test set more similar. However, we find that the detection boxes given by the ground truth of the common dataset contain a large part of the background area, the area will affect the training process as noise. To solve this problem, we propose a mask data priming method. Specifically, we redo the large scale Retail Product Checkout (RPC) dataset and add segmentation annotation information to each item in the training set image based on the original dataset using pixel-level annotation. Secondly, a new network structure is proposed in which we train the network using joint learning of detectors and counters, and fine-tune the detection network by filtering out suitable images from the test set. Experiments on the RPC dataset have shown that our method yields better results. we used an approach that reached 81.87% compared to 56.68% for the baseline approach which demonstrates that pixel-level information helps to improve the detection results of the network.

Purpose: Carotid intraplaque hemorrhage (IPH) increases risk of territorial cerebral ischemic events, but different sequences or criteria have been used to diagnose or quantify carotid IPH. The purpose of this study was to compare manual segmentation and semi-automatic segmentation for quantification of carotid IPH on magnetization-prepared rapid acquisition with gradient-echo (MPRAGE) sequences. Methods: Forty patients with 16–79% carotid stenosis and IPH on MPRAGE sequences were reviewed by two trained radiologists with more than five years of specialized experience in carotid plaque characterization with carotid plaque MRI. Initially, the radiologists manually viewed the IPH based on the MPRAGE sequence. IPH volume was then measured by three different semi-automatic methods, with high signal intensity 150%, 175%, and 200%, respectively, above that of adjacent muscle on the MPRAGE sequence. Agreement on measurements between manual segmentation and semi-automatic segmentation was assessed using the intraclass correlation coefficient (ICC). Results: There was near-perfect agreement between manual segmentation and the 150% and 175% criteria for semi-automatic segmentation in quantification of IPH volume. The ICC of each semi-automatic segmentation were as follows: 150% criteria: 0.861, 175% criteria: 0.809, 200% criteria: 0.491. The ICC value of manual vs. 150% criteria and manual vs. 175% criteria were significantly better than the manual vs. 200% criteria (p < 0.001). Conclusions: The ICC of 150% and 175% criteria for semi-automatic segmentation are more reliable for quantification of IPH volume. Semi-automatic classification tools may be beneficial in large-scale multicenter studies by reducing image analysis time and avoiding bias between human reviewers.

Ship speeds extracted from AIS data vary with ice conditions. We extrapolated this variation with SAR data to a chart of expected icegoing speed. The study is for the Gulf of Bothnia in March 2013 and for ships with ice class 1A Super that are able to navigate without icbreaker assistance. The speed was normalized to 0-10 for each ship. As the matching between AIS and SAR was complicated by ice drift during the time gap, from hours to two days, we calculated a set of local SAR statistics over several scales. We used random tree regression to estimate the speed. The accuracy was quantified by mean squared error (MSE), and the fraction of estimates close to the actual speeds. These depended strongly on the route and the day. MSE varied from 0.4 to 2.7 units² for daily routes. 65 % of the estimates deviated less than one unit and 82 % less than 1.5 units from the AIS speeds. The estimated daily mean speeds were close to the observations. Largest speed decreases were provided by the estimator in a dampened form or not at all. This improved when ice chart thickness was included as one predictor.

Background: The human mind is multimodal. Yet most behavioral studies rely on century-old measures such as task accuracy and latency. To create a better understanding of human behavior and brain functionality, we should introduce other measures and analyze behavior from various aspects. However, it is technically complex and costly to design and implement the experiments that record multiple measures. To address this issue, a platform that allows synchronizing multiple measures from human behavior is needed. Method: This paper introduces an opensource platform named OpenSync, which can be used to synchronize multiple measures in neuroscience experiments. This platform helps to automatically integrate, synchronize and record physiological measures (e.g., electroencephalogram (EEG), galvanic skin response (GSR), eye-tracking, body motion, etc.), user input response (e.g., from mouse, keyboard, joystick, etc.), and task-related information (stimulus markers). In this paper, we explain the structure and details of OpenSync, provide two case studies in PsychoPy and Unity. Comparison with existing tools: Unlike proprietary systems (e.g., iMotions), OpenSync is free and it can be used inside any opensource experiment design software (e.g., PsychoPy, OpenSesame, Unity, etc., https://pypi.org/project/OpenSync/ and https://github.com/moeinrazavi/OpenSync_Unity). Results: Our experimental results show that the OpenSync platform is able to synchronize multiple measures with microsecond resolution.

Taking into account the increase in the emission of greenhouse gases produced by ships, during navigation and maneuvering in the port, a direct consequence of the increase in maritime traffic, the international community has developed a broad set of regulations to limit such emissions. The installation in commercial ports of automatic mooring systems by means of vacuum suction cups (AMS), thus reducing considerably the time required to carry out the mooring and unmooring maneuvers of ships, is a factor that is considerably influencing the decrease in Emissions of polluting gases in commercial ports with high traffic. The objective of the present work is to verify the influence of the use of the AMS on the emissions of polluting gases produced in the facilities destined to the traffic of container ships. It examines the CO2 emissions of container ships that call in the only three container ports equipped with AMS: Salalah (Oman), Beirut (Lebanon) and Ngqura (South Africa). Between them, these three ports supported the transit of 6 million TEUS in 2017. The calculation of emissions is made taking into account the time saved when performing the mooring maneuvers using the new AMS system compared with when it is not used. To do this, two different calculation methods are used: EPA and ENTEC to then compare the results of the two and thus obtain the reduction in emissions per TEU in these terminals during the mooring maneuvers. The paper concludes with a discussion on the values of the reductions in emissions obtained and the advantages of the installation of AMS in commercial ports located near population centres.

The aim of the instant study was to evaluate relative inline progesterone dynamic changes according to parity and status of reproduction and to estimate the relationship with productivity in dairy cows by inline milk analysis system (IMAS) Herd Navigator. According to a progesterone assay, cows were divided into three periods: postpartum, after insemination, and pregnancy. In the first stage of the postpartum period (0-29 days), the progesterone level in milk was monitored every 6 days. The second stage of the postpartum period (30-65 days) lasted until cows were inseminated. In the third period (0-45 days) after cows were inseminated, progesterone scores were distributed according to whether or not cows became pregnant. The stability of progesterone dynamics was monitored in the last study period (45-90 days). For milk progesterone detection, the fully automated real-time progesterone analyzer Herd Navigator (Lattec I/S. Hillerød. Denmark) was used in combination with a DeLaval milking robot (DeLaval Inc., Tumba, Sweden). The highest progesterone concentration in multiparous cows ranged from 1.08% (11-17 days postpartum) to 34.89% higher than that in cows of the first parity. The lowest progesterone concentrations in the milk of all cows were estimated during the first 5 postpartum days and between 18 and 23 days after calving. Peak milk progesterone concentrations were evaluated in the first stage of the experiment on days 24-29 after calving. In the 30-65-day period after calving, the level of milk progesterone was 2.02-2.08 times higher than that in the 24-29-day postpartum period. After insemination, the level of progesterone in milk increased by 10.77-22.54% compared with the level from cows on days 30-65 after calving. A higher (12.88%) concentration of progesterone in milk was evaluated in multiparous cows compared with that from cows of the first parity. In pregnant cows, milk progesterone within 0-45 days after insemination was 23.88% (in multiparous cows) and 32.54% (in primiparous cows) higher than that in non-pregnant cows. On days 31–35 after insemination, pregnant cows had higher milk progesterone levels, which can predict pregnancy success. According to our study results, we can suggest that an inline progesterone concentration determined by inline milk analysis system Herd Navigator and changes in its dynamics correlate with different reproductive statuses and milk yield of cows. Pregnant cows 11–15 days after insemination have higher milk progesterone levels, what positively, associated with a successful pregnancy.

Space-borne Synthetic Aperture Radar (SAR) Interferometry (InSAR) is now a key geophysical tool for surface deformation studies. The European Commission’s Sentinel-1 Constellation began acquiring data systematically in late 2014. The data, which are free and open access, have global coverage at moderate resolution with a 6 or 12-day revisit, enabling researchers to investigate large-scale surface deformation systematically through time. However, full exploitation of the potential of Sentinel-1 requires specific processing approaches as well as the efficient use of modern computing and data storage facilities. Here we present LiCSAR, an operational system built for large-scale interferometric processing of Sentinel-1 data. LiCSAR is designed to automatically produce geocoded wrapped and unwrapped interferograms and coherence estimates, for large regions, at 0.001° resolution (WGS-84 system). The products are continuously updated in a frequency depending on prioritised regions (monthly, weekly or live update strategy). The products are open and freely accessible and downloadable through an online portal. We describe the algorithms, processing, and storage solutions implemented in LiCSAR, and show several case studies that use LiCSAR products to measure tectonic and volcanic deformation. We aim to accelerate the uptake of InSAR data by researchers as well as non-expert users by mass producing interferograms and derived products.

Nowadays, an analysis of virtual media to predict society’s reaction to any events or processes is a task of great relevance. Especially it concerns meaningful information on healthcare problems. Internet sources contain a large amount of pharmacologically meaningful information useful for pharmacovigilance purposes and repurposing drug use. An analysis of such a scale of information demands developing the methods that require the creation of a corpus with labeled relations among entities. Before, there have been no such Russian language datasets. This paper considers the first Russian language dataset where labeled entity pairs are divided into multiple contexts within a single text (by used drugs, by different users, by the cases of use, etc.), and a method based on the XLM-RoBERTa language model, previously trained on medical texts to evaluate the state-of-the-art accuracy for the task of indication of the four types of relationships among entities: ADR–Drugname, Drugname–Diseasename, Drugname–SourceInfoDrug, Diseasename–Indication. As shown based on the presented dataset from the Russian Drug Review Corpus, the developed method achieves the F1-score of 81.2% (obtained using cross-validation and averaged for the four types of relationships), which is 7.8% higher than the basic classifiers.

Considering the significant advancements in autonomous vehicle technology, research in this field is of interest to researchers. To automatic parking, controlling steer angle, gas hatch, and brakes need to be learned. Due to the increase in the number of cars and road traffic, car parking space has decreased. Its main reason is information error. Because the driver does not receive the necessary information or receives it too late, he cannot take appropriate action against it. This paper uses two phases: the first phase, for goal coordination, was used genetic algorithms and the Cuckoo search algorithm was used to increase driver information from the surroundings. Using the Cuckoo search algorithm and considering the limitations, it increases the driver’s level of information from the environment. Also, by exchanging information through the application, it enables the information to reach the driver much more quickly and the driver reacts appropriately at the right time. The suggested protocol is called the MODM-based solution. Here, the technique is assessed through extensive simulations performed in the NS-3 environment. Based on the simulation outcomes, it is indicated that the parking system performance metrics are enhanced based on the detection rate, false-negative rate, and false-positive rate.

High electricity demand for consumption at current supply level in Indonesia led to the rising cost of electricity bills. This factor is compounded by the fact that many electric generators in Indonesia still use fossil fuels, which contributes to the high basic generation cost. UBL is one of the universities that aim to be a green campus. Our research explores the possibility of installing motion sensor to contribute on the energy efficiency. Although mostly common in developed countries, the use of motion sensor for energy efficiency is still rare, especially in Indonesia. despite rising cost and supply shortages, Indonesian buildings are still of high energy consumption. our experiment shows that simple installation of commercially available motion sensors can contribute to reduce the electricity bill from the increase of energy efficiency. One of the efforts to lower energy demand in consumer side is to use the electricity efficiently, such as turning off lights in a room when it's not in use. This method can be simply done by turning the light switches for office and classrooms, but difficult to do in public spaces such as toilets and corridors. Our experiment shows that simple installation of commercially available motion sensors can contribute to reduce the electricity bill from the increase of energy efficiency. Automatic light switches experimentally installed in sample toilet room prove that electricity consumption from the lamps can contribute to the reduction of total weekly energy that translates into Green House Gas emission reduction.

Human visual system (HVM) can quickly localize the most salient object in scenes, which has been widely applied on natural image segmentation [15]-[19]. In ultrasound (US) breast images, compared with background areas, tumor is more salient because of its higher contrast. In this paper, we develop a novel automatic localization method based on HVM for automatic segmentation of ultrasound (US) breast tumors. First, the input image is smoothed by convolution with a linearly separable Gaussian filter and then subsampled into a 9-layer Gaussian pyramid. Then intensity, blackness ratio, and superpixel contrast features are combined to compute saliency map, in which Winner Take All algorithm is used to localize the most salient region, presenting with a circle on the localized target. Finally the circle is taken as the initial contour of CV level set to finish the extraction of breast tumor. The localization method has been tested on 400 US beast images, among which 378 images have higher saliency than background areas and succeed in localization, with high accuracy 92.00%. The HVM localization method can be used to localize the tumors, combined with this method, CV level set can achieve the fully automatic segmentation of US breast tumors. By combing intensity, blackness ratio and superpixel contrast features, the proposed localization method can successfully avoid the interference caused by background areas with low echo and high intensity. Moreover, multi-object localization of US breast images can be considered in future employment.

In this paper, we introduce a model for the automatic generation of literary sentences in French. It is based on algorithms that we have previously used to generate sentences in Spanish and Portuguese, and on a new corpus consisting of literary texts in French that we have constructed, called [FR]. Our automatic text generation algorithm combines language models, shallow parsing and deep learning, artificial neural networks. We have also proposed and implemented a manual evaluation protocol to assess the quality of the artificial sentences generated by our algorithm, by testing if they fulfill four simple criteria. We have obtained encouraging results from the evaluators for most of the desired features of our artificially generated sentences.

Reliable quantification of pulmonary arterial pressure is essential in the diagnostic and prognostic assessment of a range of cardiovascular pathologies including rheumatic heart disease, yet an accurate and routinely available method for its quantification remains elusive. This work proposes an approach to infer pulmonary arterial pressure based on scientific machine learning techniques and non-invasive, clinically available measurements. A 0-D multicompartment model of the cardiovascular system was optimized using several optimization algorithms, subject to forward-mode automatic differentiation. Measurement data were synthesized from known parameters to represent the healthy, mitral regurgitant, aortic stenosed and combined valvular disease situations with and without pulmonary hypertension. Eleven model parameters were selected for optimization based on 95 % explained variation in mean pulmonary arterial pressure. A hybrid Adam and limited-memory Broyden-Fletcher-Goldfarb-Shanno optimizer yielded the best results with input data including valvular flow rates, heart chamber volume changes and systematic arterial pressure. Mean absolute percentage errors ranged from 1.8 % to 3.78 % over the simulated test cases. The model was able to capture pressure dynamics under hypertensive conditions with pulmonary arterial systole, diastole, and mean pressure average percentage errors of 1.12 %, 2.49 % and 2.14 %, respectively. The relatively low errors highlight the potential of the proposed model to recover pulmonary pressures for diseased heart valve and pulmonary hypertensive conditions.

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 Medi-test system we developed was motivated by the large number of resources available for the medical domain, as well as the number of tests needed in this field (during and after the medical school) for evaluation, promotion, certification, etc. Generating questions to support learning and user interactivity has been an interesting and dynamic topic in NLP since the availability of e-book curricula and e-learning platforms. Current e-learning platforms offer increased support for student evaluation, with an emphasis in exploiting automation in both test generation and evaluation. In this context, our system is able to evaluate a student’s academic performance for the medical domain. Using as input medical reference texts and supported by a specially designed medical ontology, Medi-test generates different types of questionnaires for Romanian language. The evaluation includes 4 types of questions (multiple-choice, fill in the blanks, true/false and match), can have customizable length and difficulty and can be automatically graded. A recent extension of our system also allows for the generation of tests which include images. We evaluated our system with a local testing team, but also with a set of medicine students, and user satisfaction questionnaires showed that the system can be used to enhance learning.

Multi- and hyper-spectral, multi-angular top-of-canopy reflectance data call for an efficient retrieval system which can improve the retrieval of standard canopy parameters (as albedo, LAI, fAPAR), and exploit the information to retrieve additional parameters (e.g. leaf pigments). Furthermore consistency between the retrieved parameters and quantification of uncertainties are required for many applications. % (2) methods We present a retrieval system for canopy and sub-canopy parameters (OptiSAIL), which is based on a model comprising SAIL, PROSPECT-D (leaf properties), TARTES (snow properties), a soil model (BRDF, moisture), and a cloud contamination model. The inversion is gradient based and uses codes % created by Automatic Differentiation. The full per pixel covariance-matrix of the retrieved parameters is computed. For this demonstration, single observation data from the Sentinel-3 SY_2_SYN (synergy) product is used. The results are compared with the MODIS 4-day LAI/fPAR product and PhenoCam site photography. OptiSAIL produces generally consistent and credible results, at least matching the quality of the technically quite different MODIS product. For most of the sites, the PhenoCam images support the OptiSAIL retrievals. The system is computationally efficient with a rate of 150 pixel per second (7 millisecond per pixel) for a single thread on a current desktop CPU using observations on 26 bands. Not all of the model parameters are well determined in all situations. Significant correlations between the parameters are found, which can change sign and magnitude over time. OptiSAIL appears to meet the design goals, puts real-time processing with this kind of system into reach, seamlessly extends to hyper-spectral and multi-sensor retrievals, and promises to be a good platform for sensitivity studies. The incorporated cloud and snow detection adds to the robustness of the system.

Soil moisture (SM) products derived from passive satellite missions are playing an increasingly important role in agricultural applications, especially in crop monitoring and disaster warning. Evaluating the dependability of those products before they can be used on a large scale is crucial. In this study, we assessed the level 2 (L2) SM product from the Chinese Fengyun-3C (FY-3C) radiometer against in situ measurements collected from the Chinese Automatic Soil Moisture Observation Stations (CASMOS) during a one-year period from January 1 to December 31, 2016 in Henan, which is an agricultural province in China. Four statistical parameters were used to evaluate the products’ reliability: mean difference, root-mean-square error (RMSE), unbiased RMSE (ubRMSE), and the correlation coefficient. These statistical indicators revealed that the FY-3C L2 SM product generally did not agree with the in situ SM data from CASMOS. The time-series analysis further indicated that the correlations and estimated error were highly related to the growing periods of the crops in our study area. FY-3C L2 SM data tended to overestimate soil moisture during May, August, and September, when the crops reach their maximum vegetation density, and tended to underestimate the soil moisture content during the rest of the year. The averaged correlation coefficient between FY-3C SM and the Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index was 0.55, which demonstrates that the vegetation water content of the crops considerably influences the SM product. To improve the accuracy of the FY-3C SM product, an improved algorithm that can filter out the influences of the crops should be applied in the future.

In table tennis competitions the rule violation judgment with the greatest controversy is the height of the ball serve. This is because inaccuracy in the ball height judgment, which results in erroneous judgment, is unavoidable. Thus, we designed an automatic image judgment auxiliary system for table tennis ball height during service in this study. We used a high-speed camera to record the ball toss in the table tennis service. The designed algorithm architecture can automatically search for the ball and the position of the hand action under low light source conditions. The algorithm is mainly divided into hue-saturation-value color space processing and morphology processing using Hough transform to search for the circular ball. Experiment result shows that color segmentation can successfully and accurately determine the ball position under low light conditions. The morphology method can find the position of the hand and help determine the moment when the ball leaves the hand during the service ball toss. Finally, the actual size of the target is used to estimate the actual distance unit represented by the image pixel.

Generator automatic voltage control device (AVC) is an automatic voltage and reactive power optimization control terminal on the power plant side. Aiming at improving the performance of generator automatic voltage control device, this paper introduces a detection method of AVC substation performance based on virtual dispatching terminal environment. During the testing, the upper computer of redundant power plant is connected to the virtual dispatching terminal environment to simulate the sending instructions of AVC master station, the behavior of AVC sub stations and the excitation system of generators. The performance of AVC substation is evaluated according to the feedback adjustment results, which can effectively detect the abnormal behavior of AVC substation in power plant.

Due to resource constraints, most Diamond Open Access journals publish less than 25 articles per year, and 75% of journals are not able to provide their content in XML and HTML, primarily providing only PDFs (Bosman et al., 2021, p. 7-8). In order to keep up with larger commercial publishers, a high degree of automation and streamlining of processes is necessary. The Open Source Academic Publishing Suite (OS-APS) project funded by the German Federal Ministry of Education and Research aims to achieve this. OS-APS automatically extracts the underlying XML from Word manuscripts and offers optimization and export options in various formats (PDF, HTML, EPUB). The professional corporate design, e.g., of the PDFs, is managed automatically by using templates or creating one's own using a Template Development Kit. OS-APS will also connect to scholarly-led and community-driven publishing platforms such as Open Journal Systems (OJS), Open Monograph Press (OMP), and DSpace: the software will be able to be integrated into a wide range of publication processes, whether at small, low-resource commercial Open Access Publishers, or institutional and Diamond Open Access Publishers. References: Bosman, J., Frantsvåg, J. E., Kramer, B., Langlais, P.‑C., &amp; Proudman, V. (2021). Oa Diamond Journals Study. Part 1: Findings. https://doi.org/10.5281/zenodo.4558703

Lesion studies are crucial in establishing brain-behavior relationships, and accurately segmenting the lesion represents the first step in achieving this. Manual lesion segmentation is the gold standard for chronic strokes. However, it is labor-intensive, subject to bias, and limits sample size. Therefore, our objective is to develop an automatic segmentation algorithm for chronic stroke lesions on T1-weighted MR images. Methods: To train our model, we utilized an open-source dataset: ATLAS v2.0 (Anatomical Tracings of Lesions After Stroke). We partitioned the dataset of 655 T1 images with manual segmentation labels into five subsets and performed a 5-fold cross-validation to avoid overfitting of the model. We used a deep neural network (DNN) architecture for model training. Results: To evaluate the model performance, we used three metrics that pertain to diverse aspects of volumetric segmentation including shape, location, and size. The Dice similarity coefficient (DSC) compares the spatial overlap between manual and machine segmentation. The average DSC was 0.65 (0.61- 0.67; 95% bootstrapped CI). Average symmetric surface distance (ASSD) measures contour distances between the two segmentations. ASSD between manual and automatic segmentation was 12mm. Finally, we compared the total lesion volumes and the Pearson correlation coefficient (ρ) between the manual and automatically segmented lesion volumes, which was 0.97(p-value &lt; 0.001). Conclusions: We present the first automated segmentation model trained on a large multicentric dataset. This model will enable automated on-demand processing of MRI scans and quantitative chronic stroke lesion assessment.

The work is devoted to the study of a MEMS resonator dynamics under the action of phase-locked and automatic gain control loops. Particular attention is directed to the study of the nonlinearity factor of the resonator elastic restoring force. It was found that the determination of control system parameters based on the stability analysis of the operating resonant mode, in the general case, does not provide the required phase adjustment and stabilization of the oscillation amplitude. Stable multifrequency modes of oscillations are found, an analytical study of the mechanisms of their appearance and evolution is carried out under variation of the key parameters of the system. The real regions of the control system stable operation are determined (which do not coincide, as was found, with the regions of stability of the operating resonant mode, due to the presence of hidden attractors in the phase space of the system). A methodology has been developed for identifying such areas of stable operation. A significant complication of the structure of possible motions in the system with an increase in the Q-factor of the resonator is revealed.

The article presents a new approach to the analysis of the stability of automatic systems with discrete links.In almost all modern automatic control systems, there are links that break signals in time. These are power controlled switches - transistors or thyristors operating in a pulsed mode and digital links in regulators.Time discretization significantly affects the stability of processes in the automatic control system. The theoretical analysis of such systems is rather complicated and requires a significant change in engineering approaches to analysis. In connection with the improvement of digital controllers and a significant increase in their performance, in recent years this problem has practically not been remembered. However, its mathematical "content" has not changed since the 80s of the 20th century, when discreteness began to play a major role among the problems hindering progress in automatic control systems, in terms of the transition to digital systems.In this paper, a new approach is proposed, which consists in interpreting the sampling operation by a link with the proposed frequency characteristic, which determines the suppression of input high-frequency signals. This link greatly simplifies engineering calculations and demonstrates the new capabilities of sampling systems. These possibilities include the rational distribution of digitalization resources - the number of bits and the sampling interval between the regulator channels, depending on the frequency range of the efficiency of these channels. Theoretical statements have been verified and confirmed by simulation. It is shown how this approach makes it possible to formulate new principles of construction of seemingly well-known controllers - PID controllers and variable structure systems (VSS).

We hypothesized possibility that inline registered reticulorumen pH can be as biomarker of cows reproduction and health status. Aim of this study was to evaluate the relationship of reticulorumen pH with biomarkers from automatic milking system (AMS) and some blood parameters and determinate reticulorumen pH as biomarker of cows reproduction and health status. According to cows reproductive status the cows were classified as belonging to the following four groups: 15-30 d. postpartum; 1-34 d. after insemination; 35 d. after insemination (non-pregnant); 35 d. after insemination (pregnant). According reticulorumen pH assay experimental animals were divided into four classes: 1) pH<6.22 (5.3% of cows), 2) pH - 6.22-6.42 (42.1% of cows), 3) pH - 6.42-6.62 (21.1% of cows), 4) pH >6.62 (10.5% of cows). Rumination time, body weight, milk yield, milk fat – protein ratio, milk lactose, milk somatic cell count (SCC), milk electrical conductivity of all quarters of udder were registered with the help of Lely Astronaut® A3 milking robots. The pH, temperature of the contents of cow reticulorumens and cow activity were measured using specific smaX-tec boluses. Blood gas parameters were analyzed using a blood gas analyzer (EPOC, Canada). We found that pregnant cows has higher reticulorumen pH during insemination time, comparing with non-pregnant. Cows with lower reticulorumen pH has lowest milk fat – protein ratio, and lactose concentration, and highest SCC. Cows with lowest reticulorumen pH has lowest blood pH. With increase reticulorumen pH, increases blood potasium and hematocrit, decreases CO2, saturation and sodium.

Simulator training for image-guided surgical interventions may benefit from artificial intelligence systems that control the evolution of task skills in terms of time and precision of a trainee's performance on the basis of fully automatic feed-back systems. At the earliest stages of training, novice trainees frequently focus on getting faster at the task, and may thereby compromise the optimal evolution of the precision of their performance. For automatically guiding them towards attaining an optimal speed-accuracy trade-off, an effective control system for the reinforcement/correction of strategies must be able to exploit the right individual performance criteria in the right way, reliably detect individual performance trends at any given moment in time, and alert the trainee, as early as necessary, when to slow down and focus on precision, or when to focus on getting faster. This article addresses several aspects of this challenge for speed-accuracy controlled simulator training before any training on specific surgical tasks or clinical models should be envisaged. Analyses of individual learning curves from the simulator training sessions of novices and benchmark performance data of one expert surgeon, who had no specific training in the simulator task, validate the suggested approach.

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

In the development of the safety-critical systems, it is important to perform Failure Modes and Effects Analysis (FMEA) process to identify potential failures. However, traditional FMEA activities tend to be considered difficult and time-consuming tasks. To compensate for the difficulty of the FMEA task, various types of tools are used to increase the quality and the effectiveness of the FMEA reports. This paper explains an Automatic FMEA tool which integrates the Model-based Design (MBD), FMEA, and Simulated Fault Injection techniques in a single environment. The Automatic FMEA tool has the following advantages compared to the existing FMEA analysis tool. First, the Automatic FMEA tool automatically generates FMEA reports compared to the traditional spreadsheet-based FMEA tools. Second, the Automatic FMEA tool analyzes the causality between the failure modes and the failure effects by performing model-based fault injection simulation. In order to demonstrate the applicability of the Automatic FMEA, we used the electronic fuel injection system (EFI) Simulink model. The results of the Automatic FMEA were compared to that of the legacy FMEA.

In this paper, we present a framework that automatically labels Latent Dirichlet Allocation (LDA) generated topics using sentiment and aspect terms from COVID-19 tweets to help the end-users by minimizing the cognitive overhead of identifying key topics labels. Social media platforms especially Twitter are considered as one of the most influential sources of information for providing public opinion related to a critical situation like the COVID-19 pandemic. LDA is a popular topic modelling algorithm that extracts hidden themes of documents without assigning a specific label. Thus automatic labelling of LDA-generated topics from COVID-19 tweets is a great challenge instead of following the manual labelling approach to get an overview of wider public opinion. To overcome this problem, in this paper, we propose a framework named \texttt{SATLabel} that effectively identifies significant topic labels using \textit{top unigrams features of sentiment terms and aspect terms clusters from LDA generated topics} of COVID-19 related tweets to uncover various issues related to the COVID-19 pandemic. The experimental results show that our methodology is more effective, simpler, and traces better topic labels compare to the manual topic labelling approach.

Machine learning (ML) is increasingly useful as data grows in volume and accessibility as it can perform tasks (e.g. categorisation, decision making, anomaly detection, etc.) through experience and without explicit instruction, even when the data are too vast, complex, highly variable, full of errors to be analysed in other ways , . Thus, ML is great for natural language, images, or other complex and messy data available in large and growing volumes. Selecting a ML algorithm depends on many factors as algorithms vary in supervision needed, tolerable error levels, and ability to account for order or temporal context, among many other things. Importantly, ML methods for explicitly ordered or time-dependent data struggle with errors or data asymmetry. Most data are at least implicitly ordered, potentially allowing a hidden `arrow of time’ to affect non-temporal ML performance. This research explores the interaction of ML and implicit order by training two ML algorithms on Twitter data before performing automatic classification tasks under conditions that balance volume and complexity of data. Results show that performance was affected, suggesting that researchers should carefully consider time when selecting appropriate ML algorithms, even when time is only implicitly included.

Hopfield Neural Networks (HNNs) are recurrent neural networks used to implement associative memory. Their main feature is their ability to pattern recognition, optimization, or image segmentation. However, sometimes it is not easy to provide the users with good explanations about the results obtained with them due to mainly the large number of changes in the state of neurons (and their weights) produced during a problem of machine learning. There are currently limited techniques to visualize, verbalize, or abstract HNNs. This paper outlines how we can construct automatic video generation systems to explain their execution. This work constitutes a novel approach to get explainable artificial intelligence systems in general and HNNs in particular building on the theory of data-to-text systems and software visualization approaches. We present a complete methodology to build these kinds of systems. Software architecture is also designed, implemented, and tested. Technical details about the implementation are also detailed and explained. Finally, we apply our approach for creating a complete explainer video about the execution of HNNs on a small recognition problem.

Pilots can be one of the factors in many air traffic accidents. When one or both pilots are impaired (e.g. fatigue, drunk), disabled, capable but wrong-headed, don’t have sufficient training, distracted, miscommunicate with the air traffic controller, or follow wrong instructions from the air traffic controller, the risk of accident will increase dramatically. In some of these cases, the risk can be mitigated by using big data and machine learning. The system will collect and analyze large amount of data about the state of the aircraft, e.g., the flight path, the immediate environment around the aircraft, the weather and terrain information, and the pilots’ input to control the aircraft. Additional sensors such as eye tracking devices and biological monitor can also be added to determine the condition of the pilots. If the pilots’ input do not match proper reaction to the situation or the pilots are impaired, the learning machine will first provide an advisory to the pilots. If both pilots are impaired or incapable, a warning will be sent to the flight attendants and air traffic controllers so that they can take appropriate actions. The learning machine will be trained by both accident database and an automatic training system.

One of the critical tools for early detection and subsequent evaluation of the incidence of lung diseases is chest radiography. At a time when the speed and reliability of results, especially for COVID-19 positive patients, is important, the development of applications that would facilitate the work of untrained staff involved in the evaluation is also crucial. Our model takes the form of a simple and intuitive application, into which you only need to upload X-rays: tens or hundreds at once. In just a few seconds, the physician will determine the patient's diagnosis, including the percentage accuracy of the estimate. While the original idea was a mere binary classifier that could tell if a patient was suffering from pneumonia or not, in this paper we present a model that distinguishes between a bacterial disease, a viral infection, or a finding caused by COVID-19. The aim of this research is to demonstrate whether pneumonia can be detected or even spatially localized using a uniform, supervised classification.

This study presents a platform for ex-vivo detection of cancer nodules, addressing automation of medical diagnoses in surgery and associated histological analyses. The proposed approach takes advantage of the property of cancer to alter the mechanical and acoustical properties of tissues, because of changes in stiffness and density. A force sensor and an ultrasound probe were combined to detect such alterations during force-regulated indentations. To explore the specimens, regardless of their orientation and shape, a scanned area of the test sample was defined using shape recognition applying optical background subtraction to the images captured by a camera. The motorized platform was validated using seven phantom tissues, simulating the mechanical and acoustical properties of ex-vivo diseased tissues, including stiffer nodules that can be encountered in pathological conditions during histological analyses. Results demonstrated the platform’s ability to automatically explore and identify the inclusions in the phantom. Overall, the system was able to correctly identify up to 90.3% of the inclusions by means of stiffness in combination with ultrasound measurements, paving pathway towards robotic palpation during intraoperative examinations.

Among the possible improvements of Autonomous Emergency Braking (AEB) systems, reducing the intensity of the automatic braking process by studying the kinematics and general behavior of the pedestrian while crossing is crucial to determine the progressiveness of the braking, or replacing part of the braking process by an evasive maneuver when a collision is imminent. This paper proposes the integration of an autonomous avoidance system (Automatic Emergency Steering, AES) that acts directly on the steering system to generate an evasive maneuver and avoid a possible pedestrian collision (OPREVU-AES system), as well as the assessment of its effectiveness compared to a commercial AEB system. OPREVU and VULNEUREA are research projects in which INSIA and CEDINT have cooperated to improve driving assistance systems and the safety of pedestrians and cyclists through Virtual Reality (VR) techniques. The analysis of the kinematic and dynamic response of the OPREVU-AES system is conducted in CarSim© software. The effectiveness evaluation procedure is based on the reconstruction of a sample of road vehicle-to-pedestrian crashes (INSIA-UPM database), using the PCCrash® software, and taking as an indicator the probability of head injury severity (ISP). The results show that the AEB system would have prevented part of the collisions, especially after the incorporation of the OPREVU-AES system. In most of the cases where avoidance is not possible, a significant reduction of the ISP is achieved.

When the distance between an obstacle and a power line is less than the discharge distance, a discharge arc can be generated, resulting in interruption of power supplies. Therefore, regular safety inspections are necessary to ensure safe operations of power grids. Tall vegetation and buildings are the key factors threatening the safe operation of extra high voltage transmission lines within a power line corridor. Manual or LiDAR based-inspections are time consuming and expensive. To make safety inspections more efficient and flexible, a low-altitude unmanned aerial vehicle remote-sensing platform equipped with optical digital camera was used to inspect power line corridors. We propose a semi-patch matching algorithm based on epipolar constraints using both correlation coefficient and the shape of its curve to extract three dimensional (3D) point clouds for a power line corridor. Virtual photography was used to transform the power line direction from approximately parallel to the epipolar line to approximately perpendicular to epipolar line to improve power line measurement accuracy. The distance between the power lines and the 3D point cloud is taken as a criterion for locating obstacles within the power line corridor automatically. Experimental results show that our proposed method is a reliable, cost effective and applicable way for practical power line inspection, and can locate obstacles within the power line corridor with measurement accuracies better than ±0.5 m.

COVID-19 is a severe global problem, and AI can play a significant role in preventing losses by monitoring and detecting infected persons in early-stage. This paper aims to propose a high-speed and accurate fully-automated method to detect COVID-19 from the patient's CT scan images. We introduce a new dataset that contains 48260 CT scan images from 282 normal persons and 15589 images from 95 patients with COVID-19 infections. At the first stage, this system runs our proposed image processing algorithm to discard those CT images that inside the lung is not properly visible in them. This action helps to reduce the processing time and false detections. At the next stage, we introduce a novel method for increasing the classification accuracy of convolutional networks. We implemented our method using the ResNet50V2 network and a modified feature pyramid network alongside our designed architecture for classifying the selected CT images into COVID-19 or normal with higher accuracy than other models. After running these two phases, the system determines the condition of the patient using a selected threshold. We are the first to evaluate our system in two different ways. In the single image classification stage, our model achieved 98.49% accuracy on more than 7996 test images. At the patient identification phase, the system correctly identified almost 234 of 245 patients with high speed. We also investigate the classified images with the Grad-CAM algorithm to indicate the area of infections in images and evaluate our model classification correctness.

This article investigated the Automatic Generation Control(AGC) of multi-area multi-source interconnected systems with hydropower plants, thermal power plants, and wind energy. Adaptive Neuro-fuzzy controller integrated with the cascaded proportional-integral-derivative with filter (PIDF-PIDF) is a new cascaded controller (ANF-PIDF-PIDF) that has been presented as a secondary controller for applied hybrid power systems. The recent Skill Optimization Algorithm (SOA) is employed to optimize PIDF- PIDF controller parameter gains and the Adaptive Neuro-Fuzzy controller's inputs and output scaling factors. SOA is used to update the controller parameters with integral square error (ISE) employed as the objective function. A 1% step load disturbance was considered simultaneously in all three areas. The controller's performance is evaluated and compared with and without considering the effects of wind energy sources and non-linearity for ANF-PIDF-PIDF, PIDF-PIDF, and PIDF and it was determined that the ANF-PIDF-PIDF was the most efficient. The dynamic system performance is also compared with parallel high voltage direct current (HVDC) tie-lines. The investigation clearly shows that incorporating HVDC tie-line with multi-area, multi-source provides better dynamic performance in maximum amplitude, oscillation, and settling time. Additionally, sensitivity analysis is done and the optimum controller gains does not need to be reset to uncertain values in system loading conditions. All simulation results were evaluated using MATLAB 2016b.

Only now, four decades after passive houses were designed in the US and built in Canada, the authors are able to discuss the next generation of new and retrofitted buildings in the same way. This paper presents a universal system for different types of buildings and climates that includes construction experience and is reinforcing multi-disciplinary synergies. A next generation of technology uses adaptable indoor climate, integrates HVAC and building structure; applies field monitoring of energy and indoor environment and develops a performance model that is based on artificial neutral network. This approach, includes: 1. The self-learning ANN as a part of the building management system, 2. This management system guides energy optimization in a post-construction stage 3. A two-stage construction process for new and retrofits. The first stage prescribes investment level and optimizes performance; opposite in the second stage. This paper highlights that some elements of the proposed methodology have already been applied in the practice. It also underlines that this integrated multi-disciplinary system can be applied with a different combination of technological elements.

The human mind is multimodal. Yet most behavioral studies rely on century-old measures of behavior—task accuracy and latency (response time). Multimodal and multisensory analysis of human behavior creates a better understanding of how the mind works. The problem is that designing and implementing these experiments is technically complex and costly. This paper introduces versatile and economical means of developing multimodal-multisensory human experiments. We provide an experimental design framework that automatically integrates and synchronizes measures including electroencephalogram (EEG), galvanic skin response (GSR), eye-tracking, virtual reality (VR), body movement, mouse/cursor motion and response time. Unlike proprietary systems (e.g., iMotions), our system is free and open-source; it integrates PsychoPy, Unity and Lab Streaming Layer (LSL). The system embeds LSL inside PsychoPy/Unity for the synchronization of multiple sensory signals—gaze motion, electroencephalogram (EEG), galvanic skin response (GSR), mouse/cursor movement, and body motion—with low-cost consumer-grade devices in a simple behavioral task designed by PsychoPy and a virtual reality environment designed by Unity. This tutorial shows a step-by-step process by which a complex multimodal-multisensory experiment can be designed and implemented in a few hours. When conducting the experiment, all of the data synchronization and recoding of the data to disk will be done automatically.