Help recognizing text from image

[Will Fetherolf]

All,

I need some help extracting the text from this image.  I'm using the command line version of Tesseract from UBMannheim.  I think it's 5.2 installed.  I've tried every PSM, and nothing seems to pull it out.  If I crop off the minus sign, it works perfectly.

Any tips at all would be appreciated.

TIA!

Will

[René JM Clais]

I transform your image as a real Black and white image and it works well.

 -l eng    --psm 6 '

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/58b76c0b-2d6f-4574-a9e8-24d1a7dcc3d9n%40googlegroups.com.

[Will Fetherolf]

Do you know what threshold was used to convert the color to black on white?

I'm doing these operations through a home-grown automated test system, and I can convert the image to black and white on the fly.

[René JM Clais]

I use 3 filtering :   convert to gray scales - inverted image and last one   cv2.threshold(self.img1,127,255, cv2.THRESH_BINARY)

Does it help ? I can send you the python cv2 code

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/10608a37-af71-404c-9f3d-4526ab115a19n%40googlegroups.com.

[Will Fetherolf]

That helps a lot.  Sure, please send me the code.

[René JM Clais]

The code ! I hope it will help.

To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/706e03ab-45c5-4b3c-881e-3d5e8915062an%40googlegroups.com.