[Insight-users] Image Registration 1 Example

Karthik Krishnan Karthik.Krishnan at kitware.com
Wed Dec 1 13:31:00 EST 2004


Hi Alessi

Try going through the documentation (itk software guide: 
http://itk.org/HTML/Documentation.htm),Chapter 8. Below is a breif 
description:

A crude way of registering (rigid) would be to find all possible 
permuatations of shifts of the moving image and choose the one which has 
the lowest Mean Squares Difference (MSD) with the fixed one. Instead of 
finding all permutations, often an optimization algorithm is used. The 
example you mentioned uses an optimizer 
"itk::RegularStepGradientDescentOptimizer" and a MSD cost function. The 
lower the MSD, the better the registration.

The example finds the MSD of the moving w.r.t fixed image using 
itk::MeanSquaresImageToImageMetric and shifts the image in the opposite 
direction of the peak gradient by a certain step size (look at member 
function of itk::RegularStepGradientDescentOptimizer/BaseOptimizer). 
This is done iteratively until either 200 iterations or a step size of 
.01 is reached. (itk::RegularStepGradientDescentOptimizer halves step 
sizes if an overshoot occurs during gradient descent unlike 
itk::GradientDescentOptimizer). The linear interpolator is to evaluate 
the shifted moving image between grid points, cause shifts need not be 
integral.

hope this helps
Regards
kk


Alessi Hammilton wrote:

> Hello all,
>
> I have just begun using the ITK motion registration methods. I am 
> examining the ImageRegistration1 example from ITK v1.8. It worked just 
> fine straight from the download.
>
> I would like to understand how this example is accomplishing 
> registration. I have attempted to step through the registration 
> example function call by function call but the large hierarchy of 
> classes is a bit overwhelming for me at this point.
>
> Given two simple 1D arrays representing the fixed image pixels, and 
> the moving image pixels, can anyone describe what calculations are 
> being done per pixel to reach registration?
>
> Thanks you,
> Alessi
>
> _________________________________________________________________
> Don’t just search. Find. Check out the new MSN Search! 
> http://search.msn.click-url.com/go/onm00200636ave/direct/01/
>
> _______________________________________________
> Insight-users mailing list
> Insight-users at itk.org
> http://www.itk.org/mailman/listinfo/insight-users
>




More information about the Insight-users mailing list