[SciPy-user] 2D and 3D moments of an image (Hu, Zernike, ...)

Karl Young Karl.Young@ucsf....
Tue Feb 5 16:28:23 CST 2008


There's a very abstract definition of the usage in: Young K, Schuff N. 
Measuring structural complexity in brain images. Neuroimage Vol 39/4 pp 
1721-1730 (2008) (if you can't get a copy let me know I can send you a 
pdf). To implement the special case re. the example in that paper I've 
got some kludgy code that essentially generates the cooccurence matrix 
for contiguous linear blocks of arbitrary length in dimensions up to 4. 
But many parts of that code are unnecessarily specific to linear blocks 
(as usual I was in too much of a hurry to generate results !). I'd love 
to see a more community based effort at producing something more general 
(and efficient, robust,...). But if this is something only I'm 
interested in I'm happy to just hack something that works and not 
clutter up another package with overly general (and kludgy) code. I just 
thought there might be some texture analysis types who might be 
interested in something like that. Thanks for the response; let me know 
if you have any interest, otherwise no worries.

> 2008/2/5, Karl Young <Karl.Young@ucsf.edu <mailto:Karl.Young@ucsf.edu>>:
>     I'd be very interested in that as well - I just downloaded and played
>     around with the code that Stéfan graciously provided
>     (greycomatrix). I'd
>     be interested in seeing the code generalized along the lines described
>     (arbitrary block shapes/sizes in arbitrary dimensions (well almost
>     arbitrary - 3D and 4D would be arbitrary enough for me at the
>     moment)).
>     I was thinking about hacking it myself but I'm sure I wouldn't do
>     nearly
>     as nice a job of it as e.g. Stéfan so was glad to see there might be
>     more general interest.  I'd be willing to contribute in whatever way
>     would be useful.
> If you have an article where this is explained (and used as well), I 
> would be happy to help with it. For the moment, I only saw an almost 
> 3D cooccurence matrix (doi:10.1016/S0730-725X(03)00201-7), but not for 
> a full fledged 3D COM (with application to segmentation).
> Matthieu
> -- 
> French PhD student
> Website : http://matthieu-brucher.developpez.com/
> Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92
> LinkedIn : http://www.linkedin.com/in/matthieubrucher
>SciPy-user mailing list


Karl Young
Center for Imaging of Neurodegenerative Diseases, UCSF          
VA Medical Center (114M)              Phone:  (415) 221-4810 x3114  lab        
4150 Clement Street                   FAX:    (415) 668-2864
San Francisco, CA 94121               Email:  karl young at ucsf edu

More information about the SciPy-user mailing list