1

I am currently looking into OCR solutions to implement into our software. After reading through the messages here and trying out the products I am trying to decide between Pegasus ImageGear and Atalasoft.

We currently use Atalasoft for some other functionality. They are pretty responsive, but I have found their licensing to be a giant pain. Other than that no big complaints.

I have no experience with Pegasus ImageGear other than their SDK. The SDK was great with tons of samples and a pretty robust help library.

Can anyone relay any experiences that you may have had with either company? How is Pegasus support? Did you run into any major issues trying to implement their OCR product?

Thanks.

Aaron L.
  • 225
  • 1
  • 12

1 Answers1

3

Each of those solutions you listed uses a 3rd party OCR engine. My first concern would be testing the quality of the OCR results yourself from each engine to see if the OCR results meet up with your expectations. It depends on how much OCR processing you want to do.

You could always integrate a separate OCR library and use the above libraries for image processing and viewing only. That way you get to choose the best OCR engine for the job and you will get much better control over the OCR parameters for fine tuning.

There are expensive OCR engines out there that are not worth the money and there are good engines at reasonable prices. It just depends on how clean your images are and how complex they are. Half the problem of getting good OCR is getting a good scanned image in the first place and some proper image preprocessing.

There are other imaging libraries out there that may be worth looking into as well :-

OCR libraries

You would need to research each of these options to see if they meet your needs.

Andrew Cash
  • 2,321
  • 1
  • 17
  • 11
  • I just found this comment at http://stackoverflow.com/questions/729067/ocr-net-recommended that could be helpful - "ABBYY's component is pretty expensive. I've evaluated Pegasus ImagXpress and Atalasoft DotImage and while I've found DotImage more accurate on full-page OCR, certain small portions of text which was difficult to recognize were better read by ImagXpress. I suggest you try demo versions of both and see what fits best for your needs." – Andrew Cash Jan 12 '11 at 04:30