search

Royalvice / DocDiff

ACM Multimedia 2023: DocDiff: Document Enhancement via Residual Diffusion Models. Also contains 1597 red seals in Chinese scenes, along with their corresponding binary masks.
https://www.aibupt.com/
MIT License
223 stars 22 forks source link

watermark removal #14

Closed zairm21 closed 1 year ago

zairm21 commented 1 year ago

Hi, I tested the denoiser model on a watermarked document image(using the demo notebook) and the results show the watermark is not removed: Left to right: img, init_predict.cpu(), min_max(sampledImgs), and finalImgs

Native: native

Non-native: non

Is there a different model for watermark removal task?

Royalvice commented 1 year ago

Hi, I tested the denoiser model on a watermarked document image(using the demo notebook) and the results show the watermark is not removed: Left to right: img, init_predict.cpu(), min_max(sampledImgs), and finalImgs

Native: native

Non-native: non

Is there a different model for watermark removal task?

The model in the demo is a deblurring model, which has only been trained on deblurring datasets and is incapable of removing watermarks.

zairm21 commented 1 year ago

Are you planning to release the watermark removal model?