SSP-3D, short for Sports Shape and Pose 3D, is an evaluation dataset consisting of 311 images of athletes wearing tights, covering a variety of body shapes and poses. These images come from the Sports-1M dataset. SSP-3D is designed to serve as a baseline for body shape prediction methods. Pseudo ground truth 3D shape labels (using the SMPL body model) are obtained through multi-frame optimization, and the shapes are consistent from frame to frame.

Citation

@InProceedings{STRAPS2018BMVC, author = {Sengupta, Akash and Budvytis, Ignas and Cipolla, Roberto}, title = {Synthetic Training for Accurate 3D Human Pose and Shape Estimation in the Wild}, booktitle = {British Machine Vision Conference (BMVC)}, month = {September}, year = {2020} }