1 paper
Sen Yang, Zhibin Quan, Mu Nie +1
While CNN-based models have made remarkable progress on human pose estimation, what spatial dependencies they capture to localize keypoints remains unclear. In this work, we propos…