Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
gaowx b6159a4cfd | 1 year ago | |
---|---|---|
MindSpore | 1 year ago | |
PyTorch | 1 year ago | |
TensorFlow | 1 year ago | |
LICENSE | 1 year ago | |
No-reference Point Cloud Geometry Quality Assessment Based on Pairwise Rank Learning.pdf | 1 year ago | |
README.md | 1 year ago |
key words: point cloud, geometry quality assessment, rank learning, objective quality assessment, point cloud quality assessment
PRL-GQA is a no-reference geometry-only quality assessment method of point clouds. It leverages the pairwise rank learning to predict relative geometry quality order and exhibits competitive ranking accuracy on the proposed PRLD dataset.
1.transplant from pytorch to tensorflow and mindspore.
2.benchmark test on mindspore, tensorflow and pytorch, and compare the performance.
root
└── MindSpore: mindspore code, models included
└── TensorFlow: tensorflow code, models included
└── PyTorch: models in pytorch version, for pytorch source code, please go to: https://zhiyongsu.github.io/Project/PRLGQA.html
└── No-reference Point Cloud Geometry Quality Assessment Based on Pairwise Rank Learning.pdf: origional paper
cd MindSpore
cd TensorFlow
training & test:
python ./train_test.py
Table 1. Test on PRLD dataset
Source | Accuracy | Test Time(s) | Gpu Memory(MB) |
---|---|---|---|
Paper | 0.9449 | -- | -- |
PyTorch | 0.9260 | 624.1 | 2198 |
MindSpore | 0.9159 | 1308.1 | 4012 |
TensorFlow | 0.8924 | 1962.8 | 2692 |
@article{su2022no,
title={No-reference Point Cloud Geometry Quality Assessment Based on Pairwise Rank Learning},
author={Su, Zhiyong and Chu, Chao and Chen, Long and Li, Yong and Li, Weiqing},
journal={arXiv preprint arXiv:2211.01205},
year={2022}
}
name: Ye Hua, Gao Wenxu
email: yeh@pcl.ac.cn, gaowx@stu.pku.edu.cn
point cloud, geometry quality assessment, rank learning, objective quality assessment, point cloud quality assessment
Text Python
Apache-2.0
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》