An approach to partial occlusion using deep metric learning

<span>The human face can be used as an identification and authentication tool in biometric systems. Face recognition in forensics is a challenging task due to the presence of partial occlusion features like wearing a hat, sunglasses, scarf, and beard. In forensics, criminal identification having partial occlusion features is the most difficult task to perform. In this paper, a combination of the histogram of gradients (HOG) with Euclidean distance is proposed. Deep metric learning is the process of measuring the similarity between the samples using optimal distance metrics for learning tasks. In the proposed system, a deep metric learning technique like HOG is used to generate a 128d real feature vector. Euclidean distance is then applied between the feature vectors and a tolerance threshold is set to decide whether it is a match or mismatch. Experiments are carried out on disguised faces in the wild (DFW) dataset collected from IIIT Delhi which consists of 1000 subjects in which 600 subjects were used for testing and the remaining 400 subjects were used for training purposes. The proposed system provides a recognition accuracy of 89.8% and it outperforms compared with other existing methods.</span>

MATHEMATICAL MODEL FOR THE INVESTIGATION OF HYPOXIC STATES IN THE HEART MUSCLE AT VIRAL DAMAGE

The main complications of organism damaged by SARS-CoV-2 virus are various cardiovascular system lesions. As a result, the secondary tissue hypoxia is developed and it is relevant to search the means for hypoxic state alleviation. Mathematical modeling of this process, followed by the imitation of hypoxic states development, and subsequent correction of hypoxia at this model may be one of the directions for investigations. Aim. The purpose of this study was to construct mathematical models of functional respiratory and blood circulatory systems to simulate the partial occlusion of blood vessels during viral infection lesions and pharmacological correction of resulting hypoxic state. Methods. Methods of mathematical modeling and dynamic programming were used. Transport and mass exchange of respiratory gases in organism, partial occlusion of blood vessels and influence of antihypoxant were described by the systems of ordinary nonlinear differential equations. Results. Mathematical model of functional respiratory system was developed to simulate pharmacological correction of hypoxic states caused by the complications in courses of viral infection lesions. The model was based on the theory of functional systems by P. K. Anokhin and the assumption about the main function of respiratory system. The interactions and interrelations of individual functional systems in organism were assumed. Constituent parts of our model were the models of transport and mass exchange of respiratory gases in organism, selforganization of respiratory and blood circulatory systems, partial occlusion of blood vessels and the transport of pharmacological substance. Conclusions. The series of computational experiments for averaged person organism demonstrated the possibility of tissue hypoxia compensation using pharmacological substance with vasodilating effect, and in the case of individual data array, it may be useful for the development of strategy and tactics for individual patient medical treatment.

Robust Face Recognition Under Partial Occlusion Based on Local Generic Features

Face recognition has drawn significant attention due to its potential use in biometric authentication, surveillance, security, robotics, and so on. It is a challenging task in the field of computer vision. Although the various state-of-the-art methods of face recognition in constrained environments have achieved satisfactory results, there are still many issues which are untouched in unconstrained environments, such as partial occlusions, large pose variations, etc. In this paper, the authors have proposed an approach which utilized the local generic feature (LGF) to recognize the face in the partial occlusion by fusing features scale invariant feature transform (SIFT) and multi-block local binary pattern (MB-LBP). It also utilizes robust kernel method for classification of the query image. They have validated the effectiveness of the proposed approach on the benchmark AR face database. The experimental outcomes illustrate that the proposed approach outperformed the state-of-art methods for robust face recognition.