The paper critically reviews face recognition models that are based on deep learning, specifically security and surveillance. Existing systems are susceptible to pose variation, occlusion, low resolution and even aging, even though they perform quite well under controlled conditions. The authors make a systematic review of four state-of-the-art architectures—FaceNet, ArcFace, OpenFace and SFace—through the use of five benchmark datasets, namely LFW, CPLFW, CALFW, AgeDB-30 and QMUL-SurvFace. The measures of performance are evaluated as the area under the receiver operating characteristic (ROC-AUC), accuracy, precision and F1-score. The results reflect that FaceNet and ArcFace achieve the highest accuracy under well-lit and frontal settings; when comparing SFace, this proved to have better robustness to degraded and low-resolution surveillance images. This shows the weaknesses of traditional embedding methods because bigger data sizes reduce the performance of OpenFace with all of the datasets. These results underscore the main point of this study: a comparative study of the models in difficult real life conditions and the observation of the trade-off between generalization and specialization inherent to any models. Specifically, the ArcFace and FaceNet models are optimized to perform well in constrained settings and SFace in the wild ones. This means that the selection of models must be closely monitored with respect to deployment contexts, and future studies should focus on the study of architectures that maintain performance even with fluctuating conditions in the form of the hybrid architectures.

Stroke, a major global health concern, is characterized by sudden neurological deficits and impaired cerebral function. Advancements in technology and the integration of medical records offer opportunities to enhance stroke care and diagnosis. By mining and analyzing electronic health records, valuable insights into the interdependencies of stroke risk factors can be gained, aiding in prediction. This research provides a comprehensive review of the application of machine learning classifiers in stroke prediction, considering various techniques, features, and performance measures utilized in previous studies. The novelty of this work is to emphasize the potential of machine learning classifiers in improving stroke prediction, with a focus on feature selection, data pre-processing, and model evaluation. The aim is to shed light on the strengths and limitations of different classifiers, including Decision Trees, AdaBoost, and Gradient Boost, and their performance metrics in stroke prediction. By achieving this goal, effective stroke risk assessment models can be developed, leading to improved patient outcomes through early intervention and targeted preventive measures. The findings reveal that machine learning classifiers, particularly AdaBoost and Gradient Boost, show promising performance in stroke prediction. These classifiers demonstrate high recall rates and balanced F1 scores, indicating their efficacy in identifying individuals at risk of stroke. This research contributes to the growing body of knowledge in stroke prediction and highlights the potential of machine learning techniques in enhancing stroke care. The integration of machine learning classifiers with stroke prediction holds great promise in improving patient outcomes. By harnessing the power of electronic health records and utilizing appropriate techniques and features, healthcare providers can enhance their ability to identify and intervene in stroke cases, ultimately leading to better preventive measures and care strategies.

Population growth and expansion of urban areas increase the need for the introduction of intelligent passenger traffic monitoring systems. Accurate estimation of the number of passengers is an important condition for improving the efficiency, safety and quality of transport services. This paper proposes an approach to the automatic detection and counting of people using computer vision and deep learning methods. While YOLOv8 and DeepSORT have been widely explored individually, our contribution lies in a task-specific modification of the DeepSORT tracking pipeline, optimized for dense passenger environments, strong occlusions, and dynamic lighting, as well as in a unified architecture that integrates detection, tracking, and automatic event-log generation. Our new proprietary dataset of 4047 images and 8918 labeled objects has achieved 92% detection accuracy and 85% counting accuracy, which confirms the effectiveness of the solution. Compared to Mask R-CNN and DETR, the YOLOv8 model demonstrates an optimal balance between speed, accuracy, and computational efficiency. The results confirm that computer vision can become an efficient and scalable replacement for traditional sensory passenger counting systems. The developed architecture (YOLO + Tracking) combines recognition, tracking and counting of people into a single system that automatically generates annotated video streams and event logs. In the future, it is planned to expand the dataset, introduce support for multicamera integration, and adapt the model for embedded devices to improve the accuracy and energy efficiency of the solution in real-world conditions.

Background: The aim of the study is to model the corneal dynamic deformation under an air puff excitation. The deformation response of the cornea was modeled by using linear and nonlinear viscoelastic models. The corneal deformation responses generated from the linear and nonlinear viscoelastic model were correlated with the clinical results, which were obtained from Corneal Visualization Scheimpflug Tonometer (Corvis ST) to evaluate the comparable biomechanical parameters of the cornea. Methods: A prompt deformation occurs when the external force applied to the cornea. Then a continuous deformation follows. A simple mass, spring and dashpot system were used to model human eyeball. Results: In linear viscoelastic model, the corneal elastic stiffness commanded behavior of the corneal deformation and its maximum, when the viscous component affected for its lateral shifting and marginally alter the magnitude.Whereas, in the nonlinear viscoelastic model, the corneal material nonlinearity commanded the behavior and maximum of the corneal deformation, while the viscous component marginally contributed for its lateral shifting and demonstrated the minimum affect on the magnitude and form. A multi-objective genetic algorithm-based optimization procedure was used to identify the material properties in the nonlinear viscoelastic model for 29 eyes of 20 normal people. Conclusion: The corneal deformation response model with nonlinear viscoelastic model showed to have a better fit with the corneal dynamic deformation behavior under an air pulse excitation. The biomechanical properties of the cornea in vivo can be evaluated by using and analysing dynamic deformation of the cornea under an air puff excitation model.

Glaucoma is one of the most common eye disease in which the optic nerve gradually dies. The aim of the study is to determine the corneal biomechanical properties by modelling the corneal dynamic deformation under an air puff excitation. The corneal dynamic deformation response was modeled by using nonlinear viscoelastic model, and then benchmarked with the clinical corneal dynamic deformation profile extracted from Corneal Visualization Scheimpflug Tonometer (Corvis ST) to determine the corresponding corneal biomechanical parameters. A total of 20 normal subjects were included in this study. An optimization process was developed to determine the better combination of corneal biomechanical parameters for the normal subjects. The results showed that the model can be used to determine the corneal biomechanical properties with the corneal dynamic deformation video extracted from the Corvis ST.