Custom Face Recognition Using YOLO.V3

Face recognition technique is adopted in many applications such as surveillance systems, medical field, security, robot navigation, etc. Today because of deep learning there is a large improvement in the area of face recognition. The importance of facial recognition comes to more importance at the time of a pandemic situation, for attendance marking people monitoring, in health care, social distancing etc. Many facial recognition algorithms are available today. But the performance of these algorithms lacks many factors and which lead to poor performance in real-time applications. The most commonly used algorithms for face recognition are R-CNN, Fast R-CNN, etc. The main problems with these algorithms are they have poor processing speed; they take time to produce the output. To overcome this problem, YOLO.V3 algorithm can be used for facial recognition which produces faster output. In this paper we will study on different versions of YOLO algorithms, study on existing algorithm for face detection and recognition (R-CNN),and conclude with the result obtained for the YOLO.V3 algorithm for recognizing a custom face.