Multi-View Combined Visual Similarity Retrieval For Face Database

This work proposed a new multi-view combined visual similarity retrieval approach for face database. The search and retrieval employ a novel facial corner coding and matching approach, which delivers fast searching speed for database retrieval. The experimental results demonstrate that the facial corners can be a useful identity feature for human face database retrieval. The profile view of a face provides a complementary structure of the face that is not seen in the frontal view. The proposed multi-view modified Hausdorff distance combines the facial corners of multiple views and achieved an improved accuracy. This technique is not limited to two view combined retrieval but a generic approach for multiple view combined similarity retrieval.