How to get check page orientaion

[Gunasekaran Velu]

Hi

I have multiple page document some pages are normal page some pages or 90 degree rotated.

How i check that(90 degree) or how can i get the orientation value for particular page then only i can rotate the page for OCR process.

Looking forward your reply.

Regards

Guna

[Tom Morris]

On Monday, February 29, 2016 at 5:46:22 AM UTC-5, Gunasekaran Velu wrote:

I have multiple page document some pages are normal page some pages or 90 degree rotated.

How i check that(90 degree) or how can i get the orientation value for particular page then only i can rotate the page for OCR process.

If you use -psm 1, Tesseract will attempt to figure out the page orientation. If you use -psm 0, it'll output just the orientation information for you, but, if it gets it correct, it should be able to use it itself without you having to rotate the pages yourself.

Tom

[Gunasekaran Velu]

Hi Tom

Thanks for your information.

Now i am able to do OCR for 90 degree orientation image.

> tesseract.exe 20160226132734282-4.png 20160226132734282-4 -l eng -psm 1 hocr

But when i overlay the 20160226132734282-4.html to original pdf file for searchable pdf word not recognized.

Attached 20160226132734282-4.png image and corresponding html file and also searchable pdf(output pdf) and original pdf file.

How can i do the searchable pdf based on this png and html file. Its working for normal pages(like orientation of degree 0).

I am able to search the word in the normal pages like orientation of degree 0 but not in orientation of degree 90.

Please do the needful.

Regards

Guna

[Tom Morris]

It wouldn't surprise me if the PDF renderer can't handle documents with mixed page orientations.

Please create an issue in the issue tracker so that it gets looked at (and include your example files):

https://github.com/tesseract-ocr/tesseract/issues

Tom

[Gunasekaran Velu]

Thanks Tom.

I have fixed the issue.

Regards

Guna