Occlusion-free appearance modeling of body parts for human pose estimation

In this paper we examine efficacy of occlusion-free appearance learning for part based model. Appearance modeling with less accurate appearance data is problematic because it adversely affects entire learning process. We evaluate the effectiveness of excluding occluded body parts to be modeled for better appearance modeling process. To meet this end, We employ a simple but effective occlusion detection method. We present our approach contributes to improve the performance of human pose estimation.