Dual-Criterion Quality Loss for Blind Image Quality Assessment
You can directly replace this loss function with the original MSE loss function.
You can directly integrate the code from DCQ_Loss into your project. For convenience, we have provided the core code. Additionally, we recommend using the IQA-PyTorch codebase as a foundation since our implementation is based on it. IQA-PyTorch is a very convenient library for image quality assessment. Here is the link: IQA-PyTorch.
If you find this code useful, please cite:
bib: @inproceedings{yuan2024dual, title={Dual-Criterion Quality Loss for Blind Image Quality Assessment}, author={Yuan, Desen and Wang, Lei}, booktitle={ACM Multimedia 2024} }