Multiple Face Recognition Using Genetic Algorithm

Face recognition plays a vital role in modern biometric and surveillance applications, yet its accuracy is often reduced by challenges such as changes in illumination, facial expressions, and the presence of multiple faces in a single frame. This paper proposes a Genetic Algorithm (GA)-based approach to improve multiple face recognition by optimizing both feature selection and classifier parameters. Face embeddings are extracted using FaceNet, and GA is applied to identify the most informative subsets of features while reducing redundancy. Each candidate solution is represented as a chromosome and evaluated with a multi-objective fitness function considering accuracy, computational efficiency, and robustness to pose and occlusion variations. Experimental findings demonstrate that the GA-driven optimization outperforms conventional techniques, particularly in multi-face scenarios with complex variations. The framework is scalable and can be adapted for real-world applications such as surveillance, secure access, and intelligent authentication. Future directions include integration of deep learning classifiers with GA and deploying the system for real-time execution on edge devices.