[our code for IntrA and Modelnet40 classification is released]
3D Medical Point Transformer: Introducing Convolution to Attention Networks for Medical Point Cloud Analysis [arxiv]
Author: Jianhui Yu, Chaoyi Zhang, Heng Wang, Dingxin Zhang, Yang Song, Tiange Xiang, Dongnan Liu, Weidong Cai
- Python >=3.6
- Pytorch >= 1.4
- Packages: tqdm, sklearn, visualdl, einops, natsort
- To build the CUDA kernel for FPS:
pip install "git+git://github.com/erikwijmans/Pointnet2_PyTorch.git#egg=pointnet2_ops&subdirectory=pointnet2_ops_lib"
- State-of-the-art accuracy on IntrA classification (F1 score): 0.936
- State-of-the-art accuracy on IntrA segmentation (IoU): 94.82% on healthy vessel and 82.39% on aneurysm
- ModelNet40 classification: 93.4%
If you find our data or project useful in your research, please cite:
@article{yu20213d,
title={3D Medical Point Transformer: Introducing Convolution to Attention Networks for Medical Point Cloud Analysis},
author={Yu, Jianhui and Zhang, Chaoyi and Wang, Heng and Zhang, Dingxin and Song, Yang and Xiang, Tiange and Liu, Dongnan and Cai, Weidong},
journal={arXiv preprint arXiv:2112.04863},
year={2021}
}
Our code is based on: