sikuli-driver team mailing list archive

[RaiMan <question245951@xxxxxxxxxxxxxxxxxxxxx>]

Question #245951 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/245951

    Status: Open => Answered

RaiMan proposed the following answer:
Wether Sikuli can read text from a screen or image totally depends on
the features of the used package Tesseract.

Sikuli tries to hand over the image containing the text to Tesseract
optimised in a way (resized, grayscale), that it is likely, that
Tesseract returns the correct text. Depending on many aspects of the
font and the used font features, the result might be not useable to 100%
correct.

Standard fonts used in the web in normal font sizes (small and extra
large are critical) on high contrast, even backgrounds normally return
the best results.

You do not get the information of the font, that was used.

Later this year, I will rewrite the text features  based on the Tess4J
package, that implements a Java API to the Tesseract features, so the
C++ layer is obsolete.

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