3D Point Cloud Classification Method Based on Dynamic Coverage of Local Area

基于局部区域动态覆盖的3D点云分类方法

by Changshuo Wang*, Han Wang, Xin Ning, Weisheng Tian, and Weijun Li.

Introduction

This code is the reproduction of the pytorch 1.7+ version of 3D Point Cloud Classification Method Based on Dynamic Coverage of Local Area on the ScanObjectNN.

Install

1. clone this repo

git clone https://github.com/changshuowang/DC-CNN_ScanObjectNN.git
cd DC-CNN_ScanObjectNN

2. create a conda virtual environment and activate it

conda create -n DC-CNN python=3.7 -y
conda activate DC-CNN

3. install required libs, pytorch 1.8.1, torchvision 0.9.1, etc.

Useage

Classification ScanObjectNN

Train: The dataset will be automatically downloaded, run following command to train.

By default, it will create a fold named "log/{modelName}-{msg}-{randomseed}", which includes args.txt, best_checkpoint.pth, last_checkpoint.pth, log.txt, out.txt.

python main.py

Test: To conduct voting testing, run

python voting.py --msg demo

Contact

You are welcome to send pull requests or share some ideas with us.

contact email: wangchangshuo@semi.ac.cn.

Acknowledgment

Our implementation is mainly based on the following codebases. We gratefully thank the authors for their wonderful works.

PointMLP, EllipsoidQuery, RS-CNN, Pointnet2_PyTorch

LICENSE

DC-CNN is under the Apache-2.0 license. Please contact the authors for commercial use.