Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formats.

Istvn Srndi, Alexander Hermans, Bastian Leibe

Browse the full WACV paper archive.