1

I want to create a robust algorithm to detect and classify different types of spatially-invariant blurs in an image. How should I proceed ?

Basically there are quite a few number of possible blurs that may affect an image, motion blur, out-of-focus blur etc. I would like a workflow to how to detect and classify the type of blur in the image, for example - "application of suitable transform, histogram ... ".

Cheers.

Roy2511
  • 938
  • 1
  • 5
  • 22
  • 1
    Ask a more specific question. – neatnick Jan 03 '14 at 20:35
  • Classification of blurs after detection is not done in the literature given. @doc brown – Roy2511 Jan 03 '14 at 20:49
  • Detection is done by calculating the "degree of blur" according to metric for the related type of blur. This gives you a rating for each type between 0 and 100% for each type, letting you decide, for example, which type of blur has the highest value, or which types are over a choosen threshold. So "detection" leads you straightforward to classification, there is no real difference between the problems. – Doc Brown Jan 03 '14 at 20:59

0 Answers0