Modern video surveillance systems mainly rely on human operators to monitor and interpret the behavior of individuals in real time, which may lead to severe delays in responding to an emergency. Therefore, there is a need for continued research into the designing of interpretable and more transparent emotion recognition models that can effectively detect emotions in safety video surveillance systems. This study proposes a novel technique incorporating a straightforward model for detecting sudden changes in a person’s emotional state using low-resolution photos and video frames from surveillance cameras. The proposed technique includes a method of the geometric interpretation of facial areas to extract features of facial expression, the method of hyperplane classification for identifying emotional states in the feature vector space, and the principles of visual analytics and “human in the loop” to obtain transparent and interpretable classifiers. The experimental testing using the developed software prototype validates the scientific claims of the proposed technique. Its implementation improves the reliability of abnormal behavior detection via facial expressions by 0.91–2.20%, depending on different emotions and environmental conditions. Moreover, it decreases the error probability in identifying sudden emotional shifts by 0.23–2.21% compared to existing counterparts. Future research will aim to improve the approach quantitatively and address the limitations discussed in this paper.

Cardiovascular diseases are the leading cause of death globally, highlighting the need for accurate diagnostic tools. To address this issue, we introduce a novel approach for arrhythmia detection based on electrocardiogram (ECG) that incorporates explainable artificial intelligence through three key methods. First, we developed an enhanced R peak detection method that integrates domain-specific knowledge into the ECG, improving peak identification accuracy by accounting for the characteristic features of R peaks. Second, we proposed an arrhythmia classification method utilizing a modified convolutional neural network (CNN) architecture with additional convolutional and batch normalization layers. This model processes a triad of cardio cycles—the preceding, current, and following cycles—to capture temporal dependencies and hidden features related to arrhythmias. Third, we implemented an interpretation method that explains CNN’s decisions using clinically relevant features, making the results understandable to clinicians. Using the MIT-BIH database, our approach achieved an accuracy of 99.43%, with F1-scores approaching 100% for major arrhythmia classes. The integration of these methods enhances both the performance and transparency of arrhythmia detection systems.

The non-transparency of artificial intelligence (AI) systems, particularly in deep learning (DL), poses significant challenges to their comprehensibility and trustworthiness. This study aims to enhance the explainability of DL models through visual analytics (VA) and human-in-the-loop (HITL) principles, making these systems more transparent and understandable to end users. In this work, we propose a novel approach that utilizes a transition matrix to interpret results from DL models through more comprehensible machine learning (ML) models. The methodology involves constructing a transition matrix between the feature spaces of DL and ML models as formal and mental models, respectively, improving the explainability for classification tasks. We validated our approach with computational experiments on the MNIST, FNC-1, and Iris datasets using a qualitative and quantitative comparison criterion, that is, how different the results obtained by our approach are from the ground truth of the training and testing samples. The proposed approach significantly enhanced model clarity and understanding in the MNIST dataset, with SSIM and PSNR values of 0.697 and 17.94, respectively, showcasing high-fidelity reconstructions. Moreover, achieving an F₁m score of 77.76% and a weighted accuracy of 89.38%, our approach proved its effectiveness in stance detection with the FNC-1 dataset, complemented by its ability to explain key textual nuances. For the Iris dataset, the separating hyperplane constructed based on the proposed approach allowed for enhancing classification accuracy. Overall, using VA, HITL principles, and a transition matrix, our approach significantly improves the explainability of DL models without compromising their performance, marking a step forward in developing more transparent and trustworthy AI systems.

Information technology allowing one to implement the tasks of classification, clustering, studying the topology of the data of the information component is proposed. The multidimensional feature space is reduced to the visual presentation space to determine the information content of the data. Optimized reduction of the space dimension to two-dimensional one applying multidimensional scaling methods is used. Visual definition of grouping data allows separating areas to form. The next stage is visual limitation of categories of classes using graphic separators. To enable flexibility of nonlinear areas limitation a combination of linear ones is used, thereby forming a piecewise linear set with necessary degree of sampling. Using piecewise linear constraints allows us to implement projecting into original multidimensional feature space. Visual construction of restrictive separators makes it possible to consider tolerance fields of changing of features parameters, separation measure of classes, nonlinearity of data grouping. This is followed by reverse expansion of space with the projection of the separators into n -dimensional space with the separating hyperspace synthesis. Thus the limitative areas of hyperspace for the necessary categories of classes are formed. At the same time the visualization of classification process in a hyperspace is provided. The information technology base is the multidimensional space projection into visual (two-dimensional) space construction piecewise linear limiters of studied areas, subsequent limiters projecting into multidimensional space. Thus the information technology enables us to synthesize separating hyperplanes limiting categories of classes in multidimensional space. The technology application successive stages are described.