← Back to team overview

sikuli-driver team mailing list archive

Re: [Question #229069]: alternative OCR engine for sikuli

 

Question #229069 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/229069

RaiMan posted a new comment:
Tesseract definitely is the best available free OCR engine.

... only the current implementation in Sikuli is still rather poor in
special situations (small or non regular fonts, numbers, low contrast,
...)

... and there are no standard tuning options in current Sikuli.

the best results with version 1.0.1 you get, if
- the image (region) has high contrast
- it is a rather normal/usual font
- the font size is >12 pixels in height
- only 1 line of text

If this is not sufficient for you, you have to use a combination of
image-processing (resize to 300 dpi equivalent, optimise contrast and
sharpness) and the tesseract command line program and run the OCR in a
subprocess (... but processing time will be >1-2 seconds depending on
implementation)

It is on the list, to improve the Tesseract implementation in Sikuli,
but I guess this will not be before version 1.2

-- 
You received this question notification because you are a member of
Sikuli Drivers, which is an answer contact for Sikuli.