ACL-SPC_PyTorch

This repository contains the official code to reproduce the results from the paper:

ACL-SPC: Adaptive Closed-Loop system for Self-Supervised Point Cloud Completion (CVPR 2023)

[arXiv] [presentation]

Installation

Clone this repository into any place you want.

git clone https://github.com/Sangminhong/ACL-SPC_PyTorch.git
cd ACL-SPC_PyTorch

Dependencies

You can try downloading the environment.yml and set the conda environment.

conda env create -f environment.yml --name ACL_SPC
conda activate ACL_SPC

Or try to set the environment manually.

• Python 3.8.5
• PyTorch 1.7.1
• numpy
• h5py
• numba
• scikit-learn
• open3d
• torchsummary
• pytorch3d
• KNN-CUDA
• pykdtree
• torch_scatter

Quick Start

If you want to test your partial point cloud on pretrained model you can simply run the below command.

CUDA_VISIBLE_DEVICES=0 python demo.py --input {input_filename} --model_filename {model_filename} --fine_tune False

or if you need to finetune on your own dataset, you can simply run the below command.

CUDA_VISIBLE_DEVICES=0 python demo.py --input {input_filename} --model_filename {model_filename} --fine_tune True --dataset_directory {dataset_directory} 

Expriments

Pretrained model

Download model_best.pth from this link and save them.

NOTE: The pretrained model is updated at April. 24th 2023.

You can now go to src folder and test our ACL-SPC:

CUDA_VISIBLE_DEVICES=0 python main.py --experiment_id {experiment id} --dataset_name {dataset} --class_name {plane/car/chair/table}  

Citation

If you find our code or paper useful, please consider citing:

@inproceedings{Hong2023ACLSPC,
  title={ACL-SPC: Adaptive Closed-Loop system for Self-Supervised Point Cloud Completion},
  author={Sangmin Hong and Mohsen Yavartanoo and Reyhaneh Neshatavar and Kyoung Mu Lee},
  booktitle={IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2023}
}