We are evaluating the Cloud SDK Service with mixed results so far.

We suppose that the various levels of quality in terms of character recognition are related to the fact, that the characters on the document are not on top of a solid-white background, but rather on a 2 color background with security features that become visible in certain lighting.

To the eye however, they are easily readable and many OCR scans do recognize most of the characters.

What pre-processing could be done before we send images to the Cloud Service. E.g. would it make sense to use a library like Pixastic to improve contrast and brightness before sending an image?

Does ABBYY provide any suggestions on how to increase reliability in that scenario?

PS:

I realize there is the Imaging SDK for iOS and Android, which is not an option for us as we target a web app.

The Imaging SDK docu reads: "ABBYY Mobile Imaging SDK provides developers with intelligent tools that can analyze photographs of documents captured with mobile devices to determine if they are suitable for optical character recognition (OCR) or should be retaken. It also offers powerful image processing functions to enhance visual quality of photographed documents for better viewing and reading."

The product tour reads: "These Pre-processing features allow developers to perform: Deskewing, Automatic Page Orientation detection, Perspective Correction, Texture removal, Resolution correction "

I assume this pre-processing is all done automatically when using the Cloud Service.

Thanks a lot Stefan

asked 26 Feb '14, 20:26

Stefan's gravatar image

Stefan
34

edited 03 Mar '14, 19:38

Andrey%20Isaev's gravatar image

Andrey Isaev ♦♦
2835

Could you please share or send to CloudOCRSDK@abbyy.com the image you recognize, the settings you use and the information you need to extract?

(03 Mar '14, 20:55) Anastasia Ga... ♦♦

Thanks - will do.

(03 Mar '14, 21:00) Stefan

Anastasia - I sent the images yesterday. Please let me know if additional information is needed.

(04 Mar '14, 16:54) Stefan

Anastasia, thanks for your offer to help. However, I have sent two emails to CloudOCRSDK@abbyy.com and not heard back yet nor have I received a confirmation that you are looking into our problem. Please understand that we need to make a business decision as to whether to go forward with your service or to what extend we can use it. Thanks again.

(06 Mar '14, 08:35) Stefan

Dear Stefan, thank you for the files. We are investigating the issue and will reply today.

(06 Mar '14, 16:08) Anastasia Ga... ♦♦

Just got your email. Thanks a lot! Will reply ASAP.

(06 Mar '14, 17:50) Stefan
showing 5 of 6 show 1 more comments

You are right, the pre-processing is done automatically when using ABBBYY Cloud OCR SDK. ABBYY Mobile Imaging SDK is designed for using before offline recognition.

The Source Image Recommendations could be helpful.

link

answered 11 Mar '14, 18:36

Anastasia%20Galimova's gravatar image

Anastasia Ga... ♦♦
790112

Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Tags:

×6
×3

Asked: 26 Feb '14, 20:26

Seen: 1,441 times

Last updated: 11 Mar '14, 18:36

© 2016 ABBYY. All rights Reserved. www.ABBYY.com | Privacy Policy | Legal