Robert Sachunsky

You have been building/installing tesserocr against an outdated libtesseract, which is not installed anymore/properly. It should pull in libtesseract.so.4 (stable) or libtesseract.so.5 (development). Try removing/uninstalling your libtesseract-dev (or do `make...

> The result is different from GetTSVText That's true, but the difference is small: it's just the document "preamble" that's missing: ```C++ "level\tpage_num\tblock_num\tpar_num\tline_num\tword_" "num\tleft\ttop\twidth\theight\tconf\ttext\n" ``` You should be able to...

I also see your use of `SetRectangle` as the culprit. The API doc says: > Each SetRectangle clears the recogntion results so multiple rectangles can be recognized with the same...

@sirfz, again, the problem is already in the usage example of the current README: https://github.com/sirfz/tesserocr/blob/711cbab544dbb4bd3dcf1f13aad9d0fef20fcac7/README.rst#L181-L187

> It is always possible to create custom `list.train` and `list.eval` and use those instead of the ones created by the Makefile. It could be **documented**, though. However, there's a...

> Note that training currently requires up to 4 threads even when Tesseract was built without OpenMP: > > * training (running continuously) > * asynchronous loading of lstmf files...

Note: I have been testing whether the problem lies in `PIL.Image.save()` in `_image_buffer`: I tried setting `compress_level`, `optimize` and `dpi` for the in-memory PNG serialization. I have even set up...

I believe I can also rule out https://github.com/python-pillow/Pillow/issues/3651, because the problem remains even with Pillow 7.0 built against libjpeg9-dev (`PIL.Image.core.jpeglib_version == '9.0'`).

Also, interestingly, the thresholded image generated internally (from global Otsu binarization) does not look all that different between CLI/SetImageFilename and API/SetImage – the only directly perceptible differences are at the...

@amitdo Thanks for your proposal, but the context here is already _saving to PNG_ to **prevent** any further quality degradation on the way from `PIL.Image` to `leptonica.pix`. I did not...