search

sparkfish / shabby-pages

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
48 stars 6 forks source link

when will you add pre-trained weights? #69

Open taner45 opened 1 year ago

taner45 commented 1 year ago

trained weight:)

taner45 commented 1 year ago

trained weight:)

jboarman commented 1 year ago

Hi @taner45, we’re still in pre-print status on this paper and the associated dataset, but we are hoping to be ready to release during the summer.

Can you share more about your interests or how you plan to use the project?