Skip to content

Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification

Notifications You must be signed in to change notification settings

YangFengSEU/CEDR

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 

Repository files navigation

CEDR

This repository is for Contrastive Embedding Distribution Refinement and Entropy-Aware Attention Network (CEDR) introduced in the following paper:

"Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification"

Updates

  • 03/01/2022 The paper is currently under review, and the codes will be released in the future.
  • 06/01/2022 codes for both model.py and main.py are available now.
  • 10/01/2022 Update a pre-trained model (OA: 82.90%, mAcc: 80.60%) on ScanObjectNN via google drive.
  • 10/01/2022 Pre-trained model (OA: 93.10%, mAcc: 91.10%) on ModelNet40 is available at google drive.

Network Architecture

image

Implementation Platforms

  • Python 3.6
  • Pytorch 0.4.0 with Cuda 9.1
  • Higher Python/Pytorch/Cuda versions should also be compatible

ModelNet40 Experiment

Test the pre-trained model:

  • download ModelNet40, unzip and move modelnet40_ply_hdf5_2048 folder to ./data

  • put the pre-trained model under ./checkpoints/modelnet

  • then run (more settings can be modified in main.py):

python main.py --exp_name=gbnet_modelnet40_eval --model=gbnet --dataset=modelnet40 --eval=True --model_path=checkpoints/modelnet/gbnet_modelnet40.t7

ScanObjectNN Experiment

Test the pre-trained model:

  • download ScanObjectNN, and extract both training_objectdataset_augmentedrot_scale75.h5 and test_objectdataset_augmentedrot_scale75.h5 files to ./data
  • put the pre-trained model under ./checkpoints/gbnet_scanobjectnn
  • then run (more settings can be modified in main.py):
python main.py --exp_name=gbnet_scanobjectnn_eval --model=gbnet --dataset=ScanObjectNN --eval=True --model_path=checkpoints/gbnet_scanobjectnn/gbnet_scanobjectnn.t7

Pre-trained Models

  • Python 3.6, Pytorch 0.4.0, Cuda 9.1
  • 8 GeForce RTX 2080Ti GPUs
  • using default training settings as in main.py
Model Dataset #Points Data
Augmentation
Performance
on Test Set
Download
Link
PointNet++ ModelNet40 1024 random scaling
and translation
overall accuracy: 93.1%
average class accuracy: 91.1%
google drive
GBNet ScanObjectNN 1024 random scaling
and translation
overall accuracy: 82.9%
average class accuracy: 80.6%
google drive

Acknowledgement

The code is built on GBNet. We thank the authors for sharing the codes. We also thank the Big Data Center of Southeast University for providing the facility support on the numerical calculations in this paper.

About

Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages