Face Detection Algorithm Generation Using Artificial Neural Networks

Abstract

This study examines research and literature reviews on Deep Neural Networks. In the realm of biometrics, authentication, access control, compliance, digital cards, surveillance systems, and facial recognition (FR), the process of identifying individuals using facial imaging has a variety of practical applications. Convolutional neural networks (CNNs) have shown efficacy in facial recognition, a kind of deep networks. Certain preprocessing measures, like as sampling, must be implemented for real-time systems prior to use in CovNets. Nonetheless, whole pictures (all pixel values) are sent to Cov-Nets as input, and all processes are executed by the network (feature extraction, function filtering, training). Consequently, Cov-Nets are often challenging and time-consuming to install. Cov-Nets are in the developmental phase, exhibiting little accuracy, hence possessing significant potential for future advancement. This research presents a novel approach to use a deep neural network for face recognition. This study presents a novel method. This method utilizes extracted face characteristics instead of raw pixel values as input. This minimizes complexity and yields an accuracy of 97.05% for the Yale Faces dataset.