Assigned
Status Update
No update yet.
Comments
jm...@google.com
If you map the returned 'boundingPoly's to the coordinates on the image you will see what the Vision API is doing. It simply starts at the top left and ends at the bottom right of the image, mapping all characters detected with polygons and returning the characters as output in that order.
If you are simply relying on the returned 'description' to have an ordering of characters based on the layout of the text via columns I can submit this as a feature request. The workaround in the meantime would be to send each individual column section of the image to the Vision API to perform 'DOCUMENT_TEXT_DETECTION' on.
If you are simply relying on the returned 'description' to have an ordering of characters based on the layout of the text via columns I can submit this as a feature request. The workaround in the meantime would be to send each individual column section of the image to the Vision API to perform 'DOCUMENT_TEXT_DETECTION' on.
ak...@gmail.com
If it give a result soon, please do as you please.
It would be great if the above problem would get a solution soon.
It would be great if the above problem would get a solution soon.
jm...@google.com
I have notified our engineering team of this request to have the detected text described according to its columns. There is no ETA at this time for the implementation.
al...@gmail.com
do you have news about this feature?
i am working at this moment at app that takes a picture of a document ( hand written and not ) that extract the text in specific areas and stores it for later processing.
the pictures are being taken by normal people in daily situations, the light is not very good, so my app needs to process the image and specify the are to extract the text, but it needs to maintain the original image ( croping is not an good option )
i am working at this moment at app that takes a picture of a document ( hand written and not ) that extract the text in specific areas and stores it for later processing.
the pictures are being taken by normal people in daily situations, the light is not very good, so my app needs to process the image and specify the are to extract the text, but it needs to maintain the original image ( croping is not an good option )
an...@gmail.com
We are 6 years later.
It's still unable to work with two columns text?
Despite the claim of being powered by AI and Machine Learning
It's still unable to work with two columns text?
Despite the claim of being powered by AI and Machine Learning
Description
I'm trying to extract text from the below image which contain double column. But, the cloud ocr giving response which contains a straight line text regardless of column or spaces. Sometimes, its dividing into blocks but the text is not in the format as in the image sent. Any suggestions for such cases. If u need more info, i can provide a detailed report.
Even I tried two image sizes (1600*1200 & 1024*768) to know whether the issue is of image quality. But the result is as stated above.
Please, reply as soon as possible. A bit of urgency.
Thanking You,
Regards,
Akhil Battula.