CPPF: Towards Robust Category-Level 9D Pose Estimation in the Wild — Data

Auxiliary training and evaluation data for CPPF (CVPR 2022).

• Code: https://github.com/qq456cvb/CPPF
• Project page: https://qq456cvb.github.io/projects/cppf
• Pretrained models: https://huggingface.co/qq456cvb/CPPF

Contents

Path	Size	Description
laptop.zip	~1.1 GB	Blender physically rendered laptop images, used to train the auxiliary lid/keyboard-base segmentation network (train_laptop_aux.py). Extract under data/laptop.
nocs_seg.zip	~20 MB	Detection priors for NOCS REAL275 evaluation with instance segmentation or bounding box masks. Extract under data/nocs_seg.
sunrgbd_extra/	~4 GB	Prepared extra data for SUN RGB-D evaluation (sunrgbd_pc_bbox_votes_50k_v1_val.zip point clouds with bbox votes, plus per-category scan name lists). Place under data/sunrgbd_extra and unzip the zip files.

Usage

pip install -U "huggingface_hub[cli]"
hf download qq456cvb/CPPF --repo-type dataset --local-dir cppf_data

Citation

@inproceedings{you2022cppf,
  title={CPPF: Towards Robust Category-Level 9D Pose Estimation in the Wild},
  author={You, Yang and Shi, Ruoxi and Wang, Weiming and Lu, Cewu},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2022}
}