Tesseract error Empty page!!

[ajsaprl]

Hi, I'm trying to recognize this simple image with tesseract:

Image size: 400 × 100 pixels

Image DPI: 300 pixels/inch

Does anyone knows what makes tesseract throw error like this one?

Failed to load any lstm-specific dictionaries for lang engnumCAPS!!
Tesseract Open Source OCR Engine v4.1.1 with Leptonica
Empty page!!
Empty page!!

I already resized my image to 300 DPI and I use some traineddata such as the default eng.traineddata or digits traineddata but all of them giving the same error.

Can you help me with this one?

[Lakshay Saini]

Hello,

What command are you using to perform OCR on this image?

Regards,
Lakshay

[Raju Kulkarni]

Try psm 6 that might help.

On Tue, 25 Feb 2020, 3:32 p.m. Raju Kulkarni <kulkarn...@gmail.com wrote:

Hi,

Actually the traineddata engNUMCAPS that you are using is not providing dictionary while training so that is come as warning. And for empty page that is problem with your image try directly use tesseract to process it.

--
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/0bec2946-e902-48cb-b9b4-2d76f8b48f25%40googlegroups.com.

[ajsaprl]

it's

tesseract converted.jpg out

[ajsaprl]

Turns out I need to properly crop the image so only the number that's appear. It will not work with those line.