Robust feature set matching for partial face recognition

Various approaches have been proposed to recognize complete faces. However, few have dealt with the problem of recognizing an arbitrary partial face. In real-world scenarios, human faces might easily be occluded by other objects. This made traditional face recognition algorithms, which heavily rely on face alignment and face normalization, infeasible. In this paper, we propose a partial face recognition approach based on feature set matching, which is able to align partial face patches to gallery faces automatically and is robust to occlusions as well as illumination changes. Specifically, we exploit local features instead of holistic features and devise a Metric Learned Extended Robust Point Matching (MLERPM) algorithm to match these local feature point sets. Based on this matching approach, we defined a point set distance metric to measure the difference between two keypoint sets. Experimental results on three public databases are included to show the performance of the proposed algorithm.