Hello everyone!
I am now implementing spatial pyramid image matching (Lazebnik) and
final stage of it is the running the svm classification on the feature
points to get the correspondence between two images. I totally don't
understand how to prepare the data in suitable format to SVM and how
in general do that with pairs of feature point coordinates?
Maybe somebody did that before and have some ideas and suggestions?
Please help! thank you!!

On 2/23/2012 8:02 AM, Annette Kryhitka wrote:
> Hello everyone!
> I am now implementing spatial pyramid image matching (Lazebnik) and
> final stage of it is the running the svm classification on the feature
> points to get the correspondence between two images. I totally don't
> understand how to prepare the data in suitable format to SVM and how
> in general do that with pairs of feature point coordinates?
> Maybe somebody did that before and have some ideas and suggestions?
> Please help! thank you!!

I don't know anything about your application, but perhaps I can help you 
figure it out yourself.

SVM is a method of binary classification. Do you have a binary 
classification problem? If so, what are the two classes?

Did you look at the examples in the documentation?
http://www.mathworks.com/help/toolbox/bioinfo/ug/bs3tbev-1.html

Good luck,

Alan Weiss
MATLAB mathematical toolbox documentation