• PQA-Net
• • Environment
  • Dataset
  • Training
  • Performance comparison
  • Citation
  • contributors

PQA-Net

This is the repository of PQA-Net. The original code is implemented by Pytorch, while we provide Mindspore and Tensorflow.
Key words: point cloud quality assessment, no-reference
In this case, the original implementation of PyTorch is as follows:
https://github.com/qdushl/PQA-Net

Environment

For PQA-Net-mindspore:
ubuntu 16.04
python 3.7
mindspore 2.0, installation reference: https://www.mindspore.cn/install

For PQA-Net-tensorflow:
ubuntu 16.04
python 3.7
tensorflow-gpu 2.x

Dataset

Download the datasets from "数据集" named "distortion.zip", and then store them in the specified path

Training

For mindspore,

cd ./PQA-Net-mindspore
python MainDTLQ.py
python MainLQ.py

For tensorflow,

cd ./PQA-Net-tf
python distortion.py
python regression.py

Performance comparison

Benchmark test on mindspore, tensorflow and pytorch below
Paper:

Citation

Bibtex:
@ARTICLE{liu2021pqa,
author={Liu, Qi and Yuan, Hui and Su, Honglei and Liu, Hao and Wang, Yu and Yang, Huan and Hou, Junhui},
journal={IEEE Transactions on Circuits and Systems for Video Technology},
title={PQA-Net: Deep No Reference Point Cloud Quality Assessment via Multi-view Projection},
year={2021},
volume={},
number={},
pages={1-1},
publisher={IEEE},
doi={10.1109/TCSVT.2021.3100282}
}

contributors

name: Zhang Yongchi && Haohui Li
email: zhangych02@pcl.ac.cn