Hi HN,
I built VividManga because most general image colorization models perform poorly on manga-style black-and-white line art — they often blur lines or destroy panel structure.
VividManga focuses specifically on manga coloring, using an image-to-image workflow tuned for clean line preservation, flat color regions, and consistent shading.
The core idea is to treat manga pages differently from natural images: - prioritize line preservation over texture realism - bias the model toward flat colors and controlled gradients - keep prompts minimal to reduce style drift
Currently it supports coloring single panels or full pages, with simple tag-based control.
This is still early and opinionated, and I’d really appreciate feedback from people who work with manga/comics or have experience with diffusion models and line art preservation.
iaiuse•1h ago
Happy to answer questions about the workflow, trade-offs, and where this breaks down. This is very much an ongoing experiment, so I’m especially interested in what doesn’t work well for you.