Maximizing Efficiency in Face Recognition: A VGG16 and Max Pooling Approach

In today's identification and verification systems, a person's identity recognizes using biometric technologies like fingerprint mapping, retinal scanning, and facial recognition. The conventional approach for face recognition technology typically entails four steps: face recognition, face alignment, feature extraction, and classification. Due to its superior performance compared to conventional techniques and the ease with which feature extraction and classification combine in a single architecture, deep learning has recently grown a favor for face recognition problems. Most models use Convolutional Neural Networks (CNN) for facial recognition problems, which include multiple layers and are computationally demanding in training vast datasets. This paper proposes a CNN model based on the modified VGG16 architecture for face detection. The proposed model is compact and shows good performance with better accuracy