| Re: Image Analysis using Splines [message #47703] |
Thu, 23 February 2006 11:05 |
James Kuyper
Messages: 425
Registered: March 2000
|
Senior Member |
|
|
Stefanie Stoschitzky wrote:
> Hello,
>
> I hope that you can help me.
> For my diploma thesis I'm trying to improve MR-images. I have small
> (allready corrected) Mosaics of the image which includes different
> tissues. I now want to improve the difference between these tissues
> using B-splines, but I don't know how.
Splines are an interpolation tool; if you are comparing two images at
different resolutions, you can use splines to interpolate the
low-resolution image to the same resolution as the high-resolution
image. I don't see any other way that splines would be relevant to such
a problem, but I could be mistaken.
The /SAMPLE option of REBIN() performs nearest-neighbor interpolation,
which is eqivalent to spline interpolation using outer-product
B-splines of order 0. The default bilinear interpolation performed by
REBIN() is equivalent to using outer-product B-splines of order 1. If
REBIN() the ratio of the image resolutions is an integer, and order 1
is sufficient, then all the work has already been done for you, and you
can just concentrate on looking at the results.
Otherwise, you can do 2-D spline interpolation by performing 1-D spline
interpolation in each direction. Run SPLINE() once for each column of
your input array, giving you the columns for an intermediate array.
Then run SPLINE() once for each row of the intermediate array, giving
you the rows of the final output array.
It would be more robust to compare the images by rebinning the high
resolution image to the same resolution as the low-resolution image.
Conveniently, REBIN() can also be used for this purpose, at least if
the ratio of the image sizes is an exact integer in both dimensions;
but in this case it has nothing to do with splines..
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
