Jianyi Wang
Jianyi Wang
We adjust the degradation settings slightly to avoid the huge gap between the synthetic data and real-world data. We use our settings for synthetic test. BSRGAN-light is also adopted for...
Based on our own visualization. After all, our focus is on real-world applications. We did not fine-tune it due to the time limitation. You may have a try if you...
> Okay, thanks for your clarifications! I'll try fine-tuning RealESRGAN with the same train data and degradation as yours, and report here if there are significant differences. Thanks for your...
Hi. It seems like a problem of the env. You may have to recreate the env following the yaml file strictly.
There should be some mistakes in your config. The most possible reason is that you did not load the Autoencoder model correctly.
It seems the env is not correct. You may try reinstall the env or switch to another implementation. See #13
Hi. Our model is for image quality assessment rather than classification. So I am not sure if it is suitable to directly training it for your case. Our output is...
I did not test on an image beyond 2K. But I guess you need to resize the input to avoid too large resolution, e.g., beyond 2K. The main reason, I...
> Hi IceClear, I found the SPAQ dataset have many images that beyond 2K, such as 5488x4112, 4032x3024, 4000x3000 etc, and from your experiments the SROCC/PLCC of SPAQ was very...
Hi. I think sometimes resizing the image is necessary given that the input resolution is too large. Too large inputs may make the IQA process meaningless.