Testing PSNR and SSIM Metrics on the DND Dataset

wooseoklee4 / AP-BSN

Official PyTorch implementation of "AP-BSN: Self-Supervised Denoising for Real-World Images via Asymmetric PD and Blind-Spot Network" in CVPR 2022.

MIT License

99 stars 13 forks source link

Testing PSNR and SSIM Metrics on the DND Dataset #20

Open mhs17 opened 3 weeks ago

mhs17 commented 3 weeks ago

Hello,

I hope this message finds you well. I'm sorry for reaching out to you only now with my issue. After splitting the DND dataset, I noticed that clean images were not being generated as expected. Additionally, when testing after training, the outputs consist only of noise, denoised images, and .mat files. Could you please guide me on how to measure the PSNR and SSIM metrics for the DND dataset? Your assistance would be greatly appreciated.

Thank you very much for your time and support.

wooseoklee4 commented 3 weeks ago

Hi, Thanks for your interesting in our work

Both benchmark evaluations for SIDD and DND can only be conducted on their official websites, while we cannot access the clean GT images. To my best knowledge, evaluation on DND dataset is now available on the website (https://noise.visinf.tu-darmstadt.de/)

Thanks! Wooseok Lee