Facial Authentication with Liveness Detection on Edge Computing

Robust facial authentication with liveness detection is vital for secure access control systems. Conventional methods using remote cloud servers suffer from latency, bandwidth constraints, and privacy risks. To address these issues, we propose an edge computing solution that performs on-device facial liveness detection using an optimized convolutional neural network on low-power hardware like the Raspberry Pi. This approach eliminates the need for remote servers, providing reliable authentication. Our system also manages user data by distributing facial encoding data and user profiles between the central server and edge devices as needed, proactively removing unused data to conserve storage. This minimizes transmission overhead while ensuring edge devices have the latest user information. By performing on-device liveness detection and data management, our architecture reduces latency, bandwidth usage, and privacy risks, enabling secure and privacy-preserving facial authentication for widespread IoT deployment.