Point Transformer V3: Simpler, Faster, Stronger
Xiaoyang Wu, Li Jiang, Peng-Shuai Wang, Zhijian Liu, Xihui Liu, Yu Qiao, Wanli Ouyang, Tong He, Hengshuang Zhao · CVPR 2024 · 2024
- Status
- To read
- Topics
- backbones
- Feeds
- First paper · height sampling, P1 · ICCV 2027
What they did
Point Transformer V3 replaces precise but costly mechanisms, such as k-nearest-neighbour search, with serialized neighbourhoods, favouring simplicity and scale; it is faster and lighter than its predecessor and more accurate across 3D benchmarks.
My note
No note yet.
Code
- Repository
- Pointcept/PointTransformerV3
- Framework
- PyTorch
- Licence
- MIT
- My status
- Not tried
BibTeX
@inproceedings{wu2024-ptv3,
title = {{Point Transformer V3: Simpler, Faster, Stronger}},
author = {Xiaoyang Wu and Li Jiang and Peng-Shuai Wang and Zhijian Liu and Xihui Liu and Yu Qiao and Wanli Ouyang and Tong He and Hengshuang Zhao},
booktitle = {CVPR 2024},
year = {2024},
eprint = {2312.10035},
archivePrefix = {arXiv}
}