Support for Deskew

[mantielero]

It would be useful to be able to gather the deskew information for a page from --psm 2.

[bozhodimitrov]

Hi @mantielero, can you post an example tesseract command (and if possible - sample image)?
It will be needed for reference and testing.

[mantielero]

It should work with any sample image. The example would be:

$ tesseract image.png - --psm 2
Warning: Invalid resolution 0 dpi. Using 70 instead.
Estimating resolution as 641
Orientation: 0
WritingDirection: 0
TextlineOrder: 2
Deskew angle: 0.0015

The value to gather would be Deskew angle.

[nok]

Until now we always handle the page segmentation mode with value 0 (-psm 0 in pytesseract.py#L428). By using -psm 2 we could get the new values and transform these to the requested format.

The differences:

$ root@ee39b77b1b6a:~# tesseract ./test.png - -l eng -psm 2
Orientation: 0
WritingDirection: 0
TextlineOrder: 2
Deskew angle: -0.0038

$ root@ee39b77b1b6a:~# tesseract ./test.png - -l eng -psm 0
Page number: 0
Orientation in degrees: 0
Rotate: 0
Orientation confidence: 15.38
Script: Latin
Script confidence: 466.67