3 ms·
This is great. Did you tune the recognition models at all or is it straight tesseract plus the wrapper.
by doppenhe 10y ago
This is great. Did you tune the recognition models at all or is it straight tesseract plus the wrapper.
- sirfz 10y agoThe aim of the project was to bring most tesseract's API functionality to a pythonic interface without the need for any subprocess calls. No model tuning involved, just a wrapper :)
- zo1 10y agoGreat work, btw! Feature request: It would be great if you could also hook-in to the model training API, if any, of Tesseract. That way, you could drive the training via python rather than using Tesseract's accompanying executables.
- sirfz 10y agoIt's possible the required methods are already there but I'll look into it. Would be cool if you open an issue on GitHub. Thanks for the feedback :)