The face is the primary characteristic of the body that is responsible for the individuality of each person. The use of facial characteristics as biometrics in the creation of face recognition systems is possible. In any type of company, taking and recording attendance is the most demanding responsibility. In the traditional method of taking attendance, teachers go through the process of calling out individual pupils and noting whether or not they are present. Constructing a human face identification system based on CCTV images using a variety of feature extraction and face recognition methods is the objective of this work. Picture acquisition from a CCTV system, image preprocessing, face detection, location determination, image extraction, and recognition are the components that make up the proposed system. Specifically, an unique face alignment approach is used to pinpoint the crucial spots inside faces, and a novel deep neural network is created for deeply encoding the face areas. Both of these steps take place in the first place. TheLocal Binary Pattern Histogram (LBPH) approach is used to obtain the properties of the system. The researchers then provide a hybrid Bayesian framework and Fully Convolution Network (FCN) model for the purpose of analyzing the similarity of feature vectors and attaining very competitive face classification accuracy. When assessing the efficacy of the LBPH algorithm, several parameters, including the true positive rate, the accuracy score, the false positive rate, the sensitivity score, the specificity score, and the F1-score, are considered to be very important.

LBPH, CCTV, FCN, Face recognition, feature extraction