Face Recognition using Random Frame Selection

Face recognition has emerged as one of the crucial components in many sensitive applications. Several algorithms use different technologies for Face recognition. Additionally, new algorithms can leverage emerging technology to achieve better accuracy and processing time. In this study, we propose a novel approach that employs the random frame selection technique in conjunction with the Visual Geometric Group Face (VGG-Face) and RetinaFace algorithms for face detection and recognition. The experiment was conducted in a typical classroom environment measuring 32x25 meters, accommodating approximately 80 students, where a standard camera was positioned on the front wall. Real-time face recognition using deep convolutional neural networks was implemented to get an accuracy around 90-96% with 20-40 number of random frames. These findings underscore the system’s effectiveness in crowded settings, paving the way for broader applications in high-density environments.