Fig. 5From: 3D human pose estimation model using location-maps for distorted and disconnected images by a wearable omnidirectional cameraScheme of estimating 3D human poses using heatmaps and location-maps. The CNN regression model generates heatmaps and location-maps for each joint from an input RGB image. The 3D coordinate joint positions are estimated from their location-maps Xj, Yj, and Zj at the location of the maximum in the heatmap HjBack to article page