Abstract

Recent state-of-the-art performance on human-body pose estimation has been achieved with Deep Convolu-tional Networks (ConvNets). Traditional ConvNet architec-tures include pooling layers which reduce computational re-quirements, introduce invariance and prevent over-training. These benefits of pooling come at the cost of reduced lo-calization accuracy. We introduce a novel architecture which includes an efficient ‘position refinement ’ model that is trained to estimate the joint offset location within a small region of the image. This refinement model is jointly trained in cascade with a state-of-the-art ConvNet model [21] to achieve improved accuracy in human joint loca-tion estimation. We show that the variance of our detec-tor approaches the variance of human annotations on the FLIC [20] dataset and outperforms all existing approaches on the MPII-human-pose dataset [1]. 1.