original image


I think recognize the digits from this image would be really easy, but it just can't be recognized by tesseract and a lot of online OCR. Then I process it like this.

process image


Unfortunately it still can't be recognized by the tesseract.

My tesseract parameter:

tesseract input output digits -psm 7

That only returns a single dot .

| |

The images are too small for it to properly work. I took your images and simply scaled them up to a height of 50px (constant resolution) and ran

tesseract input output -psm 7

First Images

My output resulted in

4 5 3 6 9 8 7 2 1 0

I did the same with your single 4, and before it got a . character as you did. Resized it and it worked fine.

Depending on your circumstances you may need a command line image resizing utility such as Imagemagick to process them.

| |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.