/vt/ - Virtual Youtubers

>>43515891
Tools do exist, but mostly command line stuff as far as I know (which is how I made that one, which mostly used https://github.com/zyddnys/manga-image-translator with some minor customization on the translation side) They also have their own web-based demo at https://touhou.ai/imgtrans/ but as far as I know it only takes single images and tends to have a long queue when I checked earlier rather than batching an entire manga chapter at once.

Making it all run on my poor raspberry pi server rather than having my main PC's GPU to chew through all the processing would make it a bit more of a challenge though. I think what I would need to do is offload the processing to an old NCS2 I got from my work a year or two back but never found any use for (would be nice to actually use it for something...) but I think this would require me to convert all the ML models into a different format and then rewrite a lot of stuff to account for this so I'll need to look into how feasible it is.
>>43516029
Yeah, it can do a lot of cleaning and even does fairly well with colored/complex backgrounds sometimes, though I didn't actually make this tool and just modified it a bit to suit my needs... The TL model is also based on some other guy's project so I'd really just be integrating all this other stuff into something suitable to be hosted on my server