The implementation of the text features (find and read) is completely revised in 1.1.4 and is based on:
- Tesserract (https://github.com/tesseract-ocr/tesseract) version 3.05
- Tess4J (https://github.com/nguyenq/tess4j) version 3.5
There are surely many things that might be worth to optimise, but the text feature in SikuliX currently runs with low development priority.
So in case you have to stick with 1.1.3, if you find it is working more reliable in your use cases.
Feel free, to attach a documentation of different behaviour in 1.1.4 here in this bug though.
The implementation of the text features (find and read) is completely revised in 1.1.4 and is based on:
- Tesserract (https:/ /github. com/tesseract- ocr/tesseract) version 3.05
- Tess4J (https:/ /github. com/nguyenq/ tess4j) version 3.5
There are surely many things that might be worth to optimise, but the text feature in SikuliX currently runs with low development priority.
So in case you have to stick with 1.1.3, if you find it is working more reliable in your use cases.
Feel free, to attach a documentation of different behaviour in 1.1.4 here in this bug though.