Return Bounding Boxes

[llop00]

Hi everyone! 👋

I’ve been thinking about the best approach to retrieve bounding boxes for detected text during OCR. My initial idea is to use a fuzzy matching technique to link the extracted text back to the polygon coordinates that define each paragraph.

Does this approach sound feasible, or are there alternative methods that might be more effective? Any advice or suggestions would be greatly appreciated!

Thanks in advance