Login  Register

Re: reverse unsharp mask?

Posted by Kenneth Sloan-2 on Apr 09, 2008; 7:48pm
URL: http://imagej.273.s1.nabble.com/reverse-unsharp-mask-tp3696607p3696608.html

On Apr 9, 2008, at 9:33 AM, Michael Doube wrote:

> Hi all
>
> I'm the disappointed recipient of a bunch of CT scans of bones in  
> air, to which overzealous unsharp masking was applied (at least,  
> that's how they look) - dark black borders of air surrounding over-
> bright bone edges.
>
> Is there any way to reverse this effect to get at least partially  
> back to the original contrast?
>

Well, let's see.  Unsharp masking works by subtracting a blurred  
image from the original, to get the USM image:


     Iusm  =   I - Ib

If  you can approximage Ib, then you can add it back:

     I' = Iusm + Ib'

I imagine that blurring Iusm might give an acceptable Ib'.

Of course, there are numerous parameters to this process that you'll  
also have to estimate.

[note: instead of subtracting and adding, I generally prefer  
blending. e.g.,

      Iusm = a*I + (1-a)*Ib

where (in the case of unsharp masking) a > 1.0.

the reverse process would use 0.0 < a < 1.0, and, of course, the  
imperfect approximation of Ib by Ib']

Interesting question - I'll toss it out to my class on Thursday and  
see if anyone has a spare hour to try it out.
If one of them picks up the ball and runs with it, can I get a sample  
image to test on?

--
Kenneth Sloan                                          
[hidden email]
Computer and Information Sciences                        +1-205-934-2213
University of Alabama at Birmingham              FAX +1-205-934-5473
Birmingham, AL 35294-1170                http://KennethRSloan.com/