This document presents a novel approach for face recognition based on clustering, shape detection, and corner detection. The approach first clusters face key points and applies shape and corner detection methods to detect the face boundary and corners. It then performs both face identification and recognition on a large face database. The method achieves lower false acceptance rates, false rejection rates, and equal error rates compared to previous works, and also calculates recognition time. It provides a concise 3-sentence summary of the key aspects of the document.
Related topics: