NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:2164
Title:Domes to Drones: Self-Supervised Active Triangulation for 3D Human Pose Reconstruction

This paper presents a method based on agent to select the best view for a 3D pose estimation given 2D poses available. The proposed system is evaluated on the CMU Panoptic dataset, which contains 480 VGA camera views. The reviewers agree with the originaly of the task and the proposal and on the clearness of the presentation Rebuttal was convincing and thus also the area chair agrees for an acceptance of the paper.