DeformingThings4D is a large-scale synthetic dataset. The dataset consists of 1,972 animation sequences and 122,365 frames across 31 different animal or humanoid species categories with dense 4D annotations. The humanoid category contains 200 animations and the animal category contains 1,772 animations. This dataset aims to use data-driven methods to solve the following tasks:

• Scene Flow Estimation
• Non-rigid tracking/registration
• Shape and movement completed
• Learning rigging through observation
• General non-rigid reconstruction

Citation

@article{li20214dcomplete,
title={4dcomplete: Non-rigid motion estimation beyond the observable surface.},
author={Yang Li, Hikari Takehara, Takafumi Taketomi, Bo Zheng, and Matthias Nießner},
journal={IEEE International Conference on Computer Vision (ICCV)},
year={2021}
}