Facial Recognition from Video Using Enhanced Social Collie Optimization-Based Deep Convolutional Neural Network Technique

A key feature of video surveillance systems is face recognition, which allows the identification and verification of people who appear in scenes frequently collected by a distributed network of cameras. The scientific community is interested in recognizing individual faces in videos, in part due to the potential applications and also due to the difficulty in artificial vision algorithms. In this research main contributions of the research lie in Enhanced Social Collie Optimization based deep Convolutional Neural Network model (ESCO deep CNN) which can be formed by virtuosos of coyotes and shepherd dogs through collaboration to achieve its optimization target. The benefit of the quick learning, rediscovery of local optima, and flexibility of guard dogs is the negotiation of the requirement for increased iteration and time consumption. Coyotes, the stream team members equipped with effective hunting behavior, at the same time, they provide optimal converging, which helps do a quicker detection within the classifiers. The study under examination shows that ESCO is an efficient method to tune hyperparameters, thus it makes the classifier for deep CNNs more accurate. Further, the application of HWTP features for attribution results in a more advanced representation of original videos at high-level details, which bolsters the deep CNN's expression of granular local information. The attained accuracy, precision, recall, and F-measure of the proposed model is 91.30%, 95.08%, 95.08%, and 95.08% for the number of retrieval 500, respectively.