How to prepare data for using in 'svmclassify'?

I am working on classification of images based my features extracted from wavelet transform of my images. I have computed the mean and energy of the two class of images. For each image I have about 6 feature values. ie., mean and energy of the detail subimages of wavelet transform. Totally I have 5 images in each class. I have extracted features from all the 10 images. I donot know how to feed my features to svm for training and testing?

Answer 1

[Jeff E]

For detailed help, see the "Input Arguments" section of svmtrain: http://www.mathworks.com/help/bioinfo/ref/svmtrain.html

But translated to your case, the "Training" input will be a 10 by "about 6" matrix. Each row contains the "about 6" measurements you have made from a single image.

The "Group" input will be a 10x1 vector that identifies the class of each image in the corresponding "Training" matrix. For simplicity, you can just make this a logical vector.