I would like to ask if there are any recommended models for extracting text from bubbles in the check model for this bubble and coordinates?
Hello, I would like to ask, this model is mainly for detecting the coordinates of bubbles and text in comics, so is there a corresponding model for text extraction in comics?
I just started getting involved in this area, can you give me some guidance? Thank you
most AI models for OCR will allow you to extract text cleanly, but coordinates are often approximate. If you need optimum precision, use Google cloud vision OCR (for a fee). If you want something opensource, check out tools like EasyOCR, tesseractOCR, PaddleOCR...
Hello!
I just built a web app called EbookCC for automatic comic panel splitting and text bubble OCR. It runs a custom YOLO26n-seg engine under the hood.
I'm looking for beta testers to help me break it and give feedback. If you have some comic or manga pages you want to try, please feel free to check it out: https://ebookcc.cptd.workers.dev/.
Thanks a lot, and let me know what you think!