Secure Access: A Multimodal Authentication and Thread Detection System for Person in Residence Halls

This project enhances facial recognition accuracy using state-of-the-art machine learning and computer vision techniques. It employs Haar Cascades for real-time face detection and a ResNet-based model for precise feature extraction, improving both speed and accuracy. To tackle lighting variations, diverse poses, and facial changes, a Generative Adversarial Network (GAN) is used for data augmentation, generating diverse synthetic images to enhance model adaptability. This ensures reliable performance across different environments. While the primary focus is on facial recognition and augmentation, SHA-256 password hashing is implemented for secure authentication. By combining traditional feature extraction with deep learning advancements, this system improves recognition accuracy, robustness, and adaptability in real-world scenarios..