OCR Text recognition from picture.

[lexvi]

There was already one OCR request in closed (#71) while that is similiar to what i am thinking about, but my suggestion is still different in a way that you could if you wanted enable OCR from settings to add automatically tags labeled: "ocr-text:(insert ocr result here)" or "ocr-row1:(text), ocr-row2:(text)" to your own tags and not to public tags, this would make it simpler to find an image that you know what the text reads on it but cannot remember what other tags you might have had on the image.

or even if not add it to tags but for example a related txt file to the image where it stores the OCR results and by searching by OCR results you could find the image in question.

[imtbl]

Tagging with system:tag-management and system:notes for now since I could see this working with both.

[tomodachi94]

For now, here's a client API daemon that OCRs text and then adds it to the ocr note in the file. From there, you'd need to tidy the output and add tags; perhaps ML can help here, by identifying relevant tags and ignoring the garbage. (I didn't include autotagging due to the unpredictability of the output of the OCR.)