Generic face image quality assessment databaseDespite the fact that the study of face images is an important sub-field in computer vision research, the lack of face IQA data and models limits the precision of current IQA metrics on face image processing tasks such as face super-resolution, face enhancement, and face editing. To narrow the gap, we created the largest IQA database of human faces in-the-wild called the Generic face image quality assessment 20k database (GFIQA-20k), in which 20,000 face images were rated and ensured the diversity of the individuals depicted in highly varied circumstances. We further propose a novel deep learning model to accurately predict face image quality, which, for the first time, explores the use of generative priors for IQA. |
Cite usGFIQA-20k is freely available to the research community. If you use our database in your research, you can cite it as follows:
@article{gfiqa20k,
title={Going the Extra Mile in Face Image Quality Assessment:
A Novel Database and Model},
author={Su, Shaolin and Lin, Hanhe and Hosu, Vlad and Wiedemann,
Oliver and Sun, Jinqiu and Zhu, Yu and Liu, Hantao and Zhang,
Yanning and Saupe, Dietmar},
journal={IEEE Transactions on Multimedia},
year={2023},
publisher={IEEE}}
|
Example of sampled face images with corresponding MOS (white digits)
The proposed face IQA model
