Skip to content

Neural network-based descreening filter for scanned Anime pictures

License

Notifications You must be signed in to change notification settings

curegit/deep-descreen

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Anime Halftone Inversing

CNN solution for scanning anime prints

This project is still WIP, and its codes are extremely rough. So they might not work properly.

Aiming

When you digitize paper-printed anime arts, you have to smooth halftone dots (because of offset printing). Usually, using filter-based methods (e.g. Gaussian blur) give undesired blurry results. This project solves it by using CNN (Convolutional Neural Network) which convert dotted images into smooth image.

Here are the provisional result and comparison to the baseline (Gaussian blur).

comparison

Train

python3 -m descreening.pitch.make JapanColor2011Coated.icc images pitch_dataset

Select CUDA Device

CUDA_VISIBLE_DEVICES

About

Neural network-based descreening filter for scanned Anime pictures

Resources

License

Stars

Watchers

Forks

Releases

No releases published