0

I have been developing an OMR(Optical mark recognition) software using c#. With it, I can successfully extract OMR data from a scanned document. But problem arises when the scanned document (or image, in my case) is not exactly as how I want it to be. Conventional image scanners make it almost impossible to get an accurate scan. Sometimes, the scanned document is a bit off from the right boundary and sometimes from the left one.

If there was an easy way to somehow correct the image position by placing some kind of markers before OMR data is extracted, it would fix the problem.

Please can anyone help me? Thanks

Ashkan S
  • 10,464
  • 6
  • 51
  • 80
atif
  • 1
  • 5
  • Can you be a bit more specific? What do you mean by "not exactly how you want it to be"? Is it just positioning? Anyway, show your code as well please. – AgataB Sep 04 '16 at 10:16
  • Yes, its just positioning. If I can somehow position the scanned image correctly before OMR extraction, the results will come out 100% accurate. – atif Sep 04 '16 at 10:19
  • It is not clear to me what attempts you have made so far. [How to Ask](http://stackoverflow.com/help/how-to-ask) – AgataB Sep 04 '16 at 10:20

1 Answers1

1

Your question is a very part of researches on OMR but to give you some hints search google scholar for tilt correction. This paper also can help you (4.3.3). Also take a look at this question: Efficient ways to determine tilt of an image

Community
  • 1
  • 1
Ashkan S
  • 10,464
  • 6
  • 51
  • 80