Learning Geometry-Disentangled Representation for Complementary Understanding of 3D Object Point Cloud.

This repository is built for the paper:

Learning Geometry-Disentangled Representation for Complementary Understanding of 3D Object Point Cloud (AAAI2021) [arXiv]
by Mutian Xu*, Junhao Zhang*, Zhipeng Zhou, Mingye Xu, Xiaojuan Qi and Yu Qiao.

Overview

Geometry-Disentangled Attention Network for 3D object point cloud classification and segmentation (GDANet):

Citation

If you find the code or trained models useful, please consider citing:

@misc{xu2021learning,
  title={Learning Geometry-Disentangled Representation for Complementary Understanding of 3D Object Point Cloud}, 
  author={Mutian Xu and Junhao Zhang and Zhipeng Zhou and Mingye Xu and Xiaojuan Qi and Yu Qiao},
  year={2021},
  eprint={2012.10921},
  archivePrefix={arXiv},
  primaryClass={cs.CV}

Installation

Requirements

Linux (tested on Ubuntu 14.04/16.04)
Python 3.5+
PyTorch 1.0+

Dataset

Create the folder to symlink the data later:

mkdir -p data
Object Classification:

Download and unzip ModelNet40 (415M), then symlink the path to it as follows (you can alternatively modify the path here) :

ln -s /path to modelnet40/modelnet40_ply_hdf5_2048 data
Shape Part Segmentation:

Download and unzip ShapeNet Part (674M), then symlink the path to it as follows (you can alternatively modify the path here) :

ln -s /path to shapenet part/shapenetcore_partanno_segmentation_benchmark_v0_normal data

Usage

Object Classification on ModelNet40

Train:

python main_cls.py
Test:
- Run the voting evaluation script, after this voting you will get an accuracy of 93.8% if all things go right:
  
  python voting_eval_modelnet.py --model_path 'pretrained/GDANet_ModelNet40_93.4.t7'
- You can also directly evaluate our pretrained model without voting to get an accuracy of 93.4%:
  
  python main.py --eval True --model_path 'pretrained/GDANet_ModelNet40_93.4.t7'

Shape Part Segmentation on ShapeNet Part

Train:
- Training from scratch:
  
  python main_ptseg.py
- If you want resume training from checkpoints, specify resume in the args:
  
  python main_ptseg.py --resume True
Test:

You can choose to test the model with the best instance mIoU, class mIoU or accuracy, by specifying model_type in the args:
- python main_ptseg.py --model_type 'ins_iou' (best instance mIoU, default)
- python main_ptseg.py --model_type 'cls_iou' (best class mIoU)
- python main_ptseg.py --model_type 'acc' (best accuracy)

Other information

Please contact Mutian Xu (mino1018@outlook.com) or Junhao Zhang (junhaozhang98@gmail.com) for further discussion.

Acknowledgement

This code is is partially borrowed from DGCNN and PointNet++.

ruoshui252844/GDANet