Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
yehua 8a7633c1f1 | 1 year ago | |
---|---|---|
TensorFlow | 1 year ago | |
mindspore | 1 year ago | |
LICENSE | 1 year ago | |
README.md | 1 year ago | |
Treating Point Cloud as Moving Camera Videos A No-Reference Quality Assessment Metric.pdf | 1 year ago |
key words:point cloud quality assessment, moving camera videos, no-reference, video quality assessment
VQA_PC is a kind of point cloud quality assessment method of no reference. It has excellent performances over SJTU, WPC and LSPCQA-I datasets, benefitting from projecting the point cloud files to videos and assessing the quality via VQA methods.
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 source code: please go to: https://github.com/zzc-1998/VQA_PC
└── Treating Point Cloud as Moving Camera Videos A No-Reference Quality Assessment Metric.pdf: origional paper
cd mindspore
cd TensorFlow
training:
python ./train/train_SJTU.py
test:
python ./test/test.py
Table 1. Test on SJTU dataset
Source | SRCC | PLCC | KRCC | RMSE |
---|---|---|---|---|
Paper | 0.8509 | 0.8635 | 0.6585 | 1.1334 |
PyTorch | 0.9125 | 0.9341 | 0.7634 | 0.8364 |
MindSpore | 0.9136 | 0.9346 | 0.7619 | 0.835 |
TensorFlow | 0.8774 | 0.9002 | 0.7048 | 1.0253 |
Table 2. Test on WPC dataset
Source | SRCC | PLCC | KRCC | RMSE |
---|---|---|---|---|
Paper | 0.7968 | 0.7976 | 0.6115 | 13.6219 |
PyTorch | 0.8173 | 0.8226 | 0.631 | 12.9618 |
MindSpore | 0.8069 | 0.8085 | 0.6188 | 13.4193 |
TensorFlow | 0.8296 | 0.8313 | 0.6457 | 12.6353 |
Table 3. test time and gpu memory on TESLA T4 gpu
framework | test time(s) | GPU memory(MB) |
---|---|---|
Paper | -- | -- |
PyTorch | 29.937 | 1358 |
MindSpore | 26.405 | 3110 |
TensorFlow | 44.887 | 4732 |
@article{zhang2022treating,
title={Treating Point Cloud as Moving Camera Videos: A No-Reference Quality Assessment Metric},
author={Zhang, Zicheng and Sun, Wei and Zhu Yucheng, Min, Xiongkuo and Wu Wei, and Chen Ying, and Zhai, Guangtao},
journal={arXiv preprint arXiv:2208.14085},
year={2022}
}
name: Ye Hua
email: yeh@pcl.ac.cn
point cloud quality assessment, moving camera videos, no-reference, video quality assessment
Python CSV
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》