SAMBLE: Shape-Specific Point Cloud Sampling for an Optimal Trade-Off Between Local Detail and Global Uniformity

<p> <a href="https://arxiv.org/abs/2504.19581"> <img src="https://img.shields.io/badge/PDF-arXiv-brightgreen" /></a> <a href="https://ies.iar.kit.edu/1473_1524.php"> <img src="https://img.shields.io/badge/Author-Homepage-red" /></a> <a href="https://pytorch.org/"> <img src="https://img.shields.io/badge/Framework-PyTorch-orange" /></a> </p>

🔥 News

[Feb 2025] SAMBLE is accepted by CVPR 2025! <br> [Mar 2023] APES is selected as a Highlight by CVPR 2023! <br> [Feb 2023] APES (our former work) is accepted by CVPR 2023!

🔧 Prerequisites

Create an environment and install dependencies:

conda create -n apesv2 python=3.9 -y
conda activate apesv2
pip install -r requirements.txt

Install PyTorch and Pytorch3D:

conda install pytorch==1.11.0 torchvision==0.12.0 cudatoolkit=11.3 -c pytorch -y
conda install -c fvcore -c iopath -c conda-forge fvcore=0.1.5 iopath=0.1.9 -y
conda install -c pytorch3d pytorch3d=0.7.0 -y

📦 How to run

Classification

# train:
python train_modelnet.py datasets=modelnet usr_config=YOUR/USR/CONFIG/PATH train.ddp.which_gpu=[0,1] train.epochs=200
# test:
python test_modelnet.py datasets=modelnet usr_config=YOUR/USR/CONFIG/PATH train.ddp.which_gpu=[0,1] train.epochs=200

Segmentation

# train:
python train_shapenet.py datasets=shapenet_AnTao350M usr_config=YOUR/USR/CONFIG/PATH test.ddp.which_gpu=[0,1]
# test:
python test_shapenet.py datasets=shapenet_AnTao350M usr_config=YOUR/USR/CONFIG/PATH test.ddp.which_gpu=[0,1]

📖 Citation

If you are interested in this work, please cite as below:

@inproceedings{wu_2025_samble,
author={Wu, Chengzhi and Wan, Yuxin and Fu, Hao and Pfrommer, Julius and Zhong, Zeyun and Zheng, Junwei and Zhang, Jiaming and Beyerer, J\"urgen},
title={SAMBLE: Shape-Specific Point Cloud Sampling for an Optimal Trade-Off Between Local Detail and Global Uniformity},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2025}
}

@inproceedings{wu_2023_attention,
author={Wu, Chengzhi and Zheng, Junwei and Pfrommer, Julius and Beyerer, J\"urgen},
title={Attention-Based Point Cloud Edge Sampling},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2023}
}