Chaofeng Chen

This is expected because you are using full-reference metrics: `psnr`, `lpips`. Full-reference metrics require a high quality reference image, and measures the difference to the reference. The distance between the...

@shshojaei That is ok for pieapp because it used an extra regression layer for the final results. And it makes pieapp output higher results for good image. There is no...

@nkjulia You may refer to the benchmark results under tests dir for performance of these metrics under different benchmarks and choose the one you need. I would recommend clipiqa+ if...

Sorry that I am not available to write a comprehensive summary because there are too many related papers. I have listed these works here: https://github.com/chaofengc/Awesome-Image-Quality-Assessment You may refer to specific...

The deep learning metrics are closely related with the training dataset. The `AVA` dataset is the main aesthetic dataset now. And our toolbox has some models trained on `AVA`. You...

Thanks to your interest! By default, SSIM uses YCbCr space and PSNR uses RGB space. Alternatively, you can pass the `create_metric('psnr', test_y_channel=True)` to control it.

FID是单独处理的，不受metric_mode影响。请参考forward 函数的参数列表，分别传入参数`fdir1`和`fdir2` https://github.com/chaofengc/IQA-PyTorch/blob/b7d19eaae1d64b335d5ca1f86b4c5e0e854ba50f/pyiqa/archs/fid_arch.py#L267-L279

`fdir1`和`fdir2`分别是生成的文件夹以及ground truth文件夹，即两个数据分布，FID计算两个分布之间的距离

Unfortunately, there isn't a model specifically designed for document images, and there is also a lack of datasets tailored to this task.

Thank you for the information. I'll look at this metric and consider adding it when I have the time.