ShabbyPages is a state-of-the-art corpus of born-digital document images with both ground truth and distorted versions appropriate for use in training models to reverse distortions and recover to original denoised documents.
MIT License
50
stars
6
forks
source link
fix(ci): checkout the repo before trying to tweet #26