How to create a dataset of my own to be used in nprtool (neural network toolbox) ?

I have 5 classes of images each contains at least 10 images, I extract 9 features from this images and represent features of each class of images in a matrix (example: 10x9 matrix - class1).. Still in variables in my script of the GUI program of my own too.

So how could I input this dataset to the neural network toolbox and to be used in the nprtool?

Also I'm sorry but I really didn't get it, Is it used for clustering so it clasture all the dataset into classes after all or classifier so if I trained it and input a new image features it classify it according to its class.

Thank you in advance.

Answer 1

[Greg Heath]

For classification use patternnet with

 [ I N ] = size(input)  % [ 9 50 ]
 [ O N ] = size(target) % [ 5 50 ]

% Columns of target are columns of eye(5); e.g., repmat(eye(5),1,10)

For documentation

 help patternnet

 doc patternnet

Hope this helps.

Thank you for formally accepting my answer

Greg