sikuli-driver team mailing list archive

Thread
Date

Re: [Question #258158]: find text from image returns garbage values

To: sikuli-driver@xxxxxxxxxxxxxxxxxxx
From: RaiMan <question258158@xxxxxxxxxxxxxxxxxxxxx>
Date: Mon, 24 Nov 2014 08:31:37 -0000
Reply-to: question258158@xxxxxxxxxxxxxxxxxxxxx
Sender: bounces@xxxxxxxxxxxxx

Question #258158 on Sikuli changed:
https://answers.launchpad.net/sikuli/+question/258158

    Status: Open => Answered

RaiMan proposed the following answer:
Sorry no.

The current implementation of the text() feature based on Tesseract has many limitations.
In this case it seems to be the font, that cannot be read correctly.

You might experiment with the used region, from which the text is read.

version 1.1.0 allows to directly read text from a given image:
text = Image.text(c)
where c is the image filename from your above snippet in find().
This would allow, to preprocess the given image to make it more suitable for text reading with Tesseract.

For version 1.2 it is planned, to improve the text feature - so for the
next months you have to live with this situation.

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