Synthetically occluded face recognition using local complex binary pattern

摘要

In real world applications of face recognition system occlusion is always a big challenge. This paper presents an approach using hybrid combination of DT-CWT and LBP for feature extraction to deal with various occlusion. The DT-CWT focuses on global scale invariant feature however LBP emphasis on local structure of face image .The proposed approach are tested for synthetically created occlusion at eyes, mouth, upper face ,lower face, which are general occlusion that is encountered in real world applications. The experimental result of recognition shows the robustness and effectiveness of proposed local complex binary pattern (LCBP) under different occlusion.