mimicat3d/PokeAnimDB

[CVPR 2026] 😸 MimiCAT: Mimic with Correspondence-Aware Cascade-Transformer for Category-Free 3D Pose Transfer

📦 PokeAnimDB

PokeAnimDB is a motion dataset for research on 3D character animation, motion analysis, and category-free pose transfer. It contains released character metadata and motion sequences stored in .npz format, and is used in our CVPR 2026 paper.

The released version contains 900+ characters and 32k+ motion clips.

🚀 Quick Start

This release is packaged as a compressed archive:

• data.tar.gz: dataset archive containing all character folders and motion files

Extract it from the PokeAnimDB/ directory with:

tar -xzf data.tar.gz

After extraction, the dataset root will contain a data/ directory with one folder per character. Each folder includes a rest-pose .npz file for the character and a motion/ subdirectory with motion sequence .npz files.

📁 Dataset Structure

PokeAnimDB/
├── data/
│   ├── pm0001_00_00/
│   │   ├── pm0001_00_00.npz
│   │   └── motion/
│   │       ├── attack01_0.npz
│   │       ├── ...
│   │       └── walk01_29.npz
│   └── ...
└── README.md

Each character directory contains:

• <character_id>.npz: rest-pose character metadata
• motion/*.npz: motion sequences with pose parameters and global translations

📜 License

This dataset is released for research and non-commercial use only under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).

The license applies only to the dataset organization, preprocessing, annotations, metadata, and converted motion files provided in this repository.

Pokémon and related character names, models, animations, and trademarks are the property of their respective owners. This project is an unofficial research dataset and is not affiliated with, endorsed by, or sponsored by Nintendo, Game Freak, Creatures Inc., or The Pokémon Company.

Users are responsible for ensuring that their use of this dataset complies with the licenses and terms of the original data sources.

📖 Citation

If you find our work or dataset useful, please consider citing:

@inproceedings{chai2026mimicat,
  author    = {Chai, Zenghao and Tang, Chen and Wong, Yongkang and Yang, Xulei and Kankanhalli, Mohan},
  title     = {MimiCAT: Mimic with Correspondence-Aware Cascade-Transformer for Category-Free 3D Pose Transfer},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026}
}