Faculty Facial Recognition Using Convolutional Neural Network a Tool for Smart Academic Monitoring

Faculty Academic Monitoring is one thing a university has to perform to ensure quality delivery of instructions. That includes proper attendance monitoring of classes, faculty consultations, and others. To address attendance monitoring in CCIS - College of Computing and Information Sciences, a facial detection and recognition was developed to identify a faculty face's important feature from a real-time captured image. The study uses a pre-trained model called facenet but acquires new training data sets to generate the desired classifier for a localized recognition system. The data are cleaned by eliminating duplicates, resizing, cropping into the desired dimension, and labeling each image data. The system used image pre-processing techniques like face alignment algorithm, landmarks, face detection, and local binary pattern. The study achieved a validation rate of 96.45% using the test validation function and an accuracy rate of 90% in the actual testing.