Official repo for "A Reduced-Reference Quality Assessment Metric for Textured Mesh Digital Humans", ICASSP 2024.
You can use get_featrures.py to extract the features of the textured mesh digital human in the sample_dh folder.
We also provide all the extracted features in the features folder for both reference and distorted digital humans from the SJTU-H3D and DHHQA databases.
You can use the performance.py to quickly reproduce the experimental results in the paper.
The database can be accessed from SJTU-H3D and DHHQA.
If you find our paper useful, please cite our paper as:
title={A Reduced-Reference Quality Assessment Metric for Textured Mesh Digital Humans},
author={Zhang, Zicheng and Zhou, Yingjie and Li, Chunyi and Fu, Kang and Sun, Wei and Liu, Xiaohong and Min, Xiongkuo and Zhai, Guangtao},
booktitle={2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},
year={2024},
organization={IEEE}
}
'''