KUR-creative/SickZil-Machine
A deep learning tool that automatically detects and erases text bubbles from manga and comics images for scanlation workflows.
Not currently ranked — collecting fresh signals.
star history
SickZil-Machine automates text removal during manga/comics translation by applying two deep learning models. SegNet, built on U-net architecture, performs semantic segmentation to identify text regions in manga images. ComplNet uses the Deepfill v2 inpainting network to naturally fill erased areas to match the surrounding background. Both detection and removal are fully automatic with no additional human intervention required, though users can optionally specify custom text regions.
Frequently asked
- What is KUR-creative/SickZil-Machine?
- A deep learning tool that automatically detects and erases text bubbles from manga and comics images for scanlation workflows.
- Is SickZil-Machine open source?
- Yes — KUR-creative/SickZil-Machine is open source, released under the AGPL-3.0 license.
- What language is SickZil-Machine written in?
- KUR-creative/SickZil-Machine is primarily written in Python.
- How popular is SickZil-Machine?
- KUR-creative/SickZil-Machine has 1.5k stars on GitHub.
- Where can I find SickZil-Machine?
- KUR-creative/SickZil-Machine is on GitHub at https://github.com/KUR-creative/SickZil-Machine.