Comparative Analysis of Transfer Learning CNN for Face Recognition

The practice of identifying people by looking at their faces is known as face recognition (FR). This technology is frequently used in biometric authentication, surveillance technologies, security systems, law enforcement, real-time attendance systems, smart cards, and other applications. The face recognition technology works in two stages. First, a method for picking up or extracting face characteristics is applied, followed by pattern categorization. Deep learning has lately made important contributions to face recognition technology, particularly the convolutional neural network (CNN). Deep learning employs numerous processing layers to develop data representations with varying degrees of feature extraction. Since the achievements of DeepFace and DeepID, this developing technology has altered the study landscape of facial recognition. Deep learning has dramatically improved cutting-edge performance and aided in the development of effective real-world applications. This study examines the performance of three among the most popular CNN architectures for face recognition. In the proposed work, transfer learning is used to deploy pre-trained CNN models for face recognition such as VGG16, ResNet-50, and MobileNet. Training and validation accuracy and loss were utilized as criterion to enhance the performance of the CNN algorithm. 5 Celebrity Faces Dataset from Kaggle, as well as a local dataset, were used to train and test the models. Face recognition was implemented in two ways: static and webcam-based. The models developed for this research can be used in real-time attendance and surveillance systems.