[2.0.5] Windows: OCR Change of behaviour

Asked by Sergio López Sánchez on 2021-04-29

I used to use OCR to read text from screen on Windows 10 with SikuliX 2.0.4. I use files for Spanish language.
In addition, I have also installed the last VC Redistributable for 2015+.

However, I've realized that upgrading Sikulix to 2.0.5 and keeping the same files of Tesseract, the behaviour of the text recognition has changed.

Before (v. 2.0.4), the regions with no text were read usually with a 'o' or '0' character. But now, with version 2.0.5, these same cases are recognized with different characters ('ÓN', 'ÁON'...).

Has there been a change recently in Sikuli code regarding this issue of OCR?

As a consequence, I've decided to downgrade to 2.0.4 because this unsteady , different behaviour doesn't let work properly.

Moreover, I must say that with last version of Sikuli, Eclipse doesn't get the modules/libraries from the sikuliapi.jar file ("from sikuli import *"). I must unzip the .jar and import the folder to Eclipse working directory in order to check this issue and import Sikuli libraries properly.

Question information

Language:
English Edit question
Status:
Open
For:
SikuliX Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:

Can you help with this problem?

Provide an answer of your own, or ask Sergio López Sánchez for more information if necessary.

To post a message you must log in.