Image problem

1. Dec 31, 2013

sigh1342

1. The problem statement, all variables and given/known data
I found that the blurred image is always presented by
$$A=Ku+f$$, where u is the perfect image source, and the K is transformation (blurring, sampling)
and f is the noising . The question I want to know is how can we find such K, or f when we blurred the image , for example ,using the gaussian blurring.
In other word, I want to know how can I find K or f s.t. I can do the same effect as using the
imnoise function in matlab as gaussian blurring or ...

2. Relevant equations
I found that there are matrix related to it, where the matrix is given by h=fspecial('gaussian',256);
and it is all zeros.

3. The attempt at a solution

2. Jan 1, 2014

Simon Bridge

It's an inverse problem ... you need to use some regularization approach.
http://home.comcast.net/~szemengtan/InverseProblems/chap3.pdf [Broken]

Last edited by a moderator: May 6, 2017
3. Jan 1, 2014

sigh1342

um,, I just want to know how can I find such K and f...
Thanks you.:)

4. Jan 2, 2014

Simon Bridge

And I told you - use a regularization technique.
I'm not going to write out the 10-20 pages of theory here - hence the link.