Domes to Drones: Self-Supervised Active Triangulation for 3D Human Pose Reconstruction Supplementary Material

摘要

This supplementary provides more insight into our ACTOR model and experimental setup. Section §1 describes the details of the network architecture, implementation, and hyperparameters. §2 elaborates on how we match 2d pose estimates in space and time using instance features. In §3 we provide 2d reprojection errors onto 2d OpenPose [2] estimates on the Panoptic test splits. Finally, §4 describes further dataset details.