Classification with Neural Network Toolbox

I attempt to build a neural network that can classify the health of leafs. I created an input Matrix with the attributes (reflection of wavelength) and a Output Matrix which contains the classes (each column is a 2-element class vector with a 1 in the position of the associated health). Those are just sample data, which I created manuelly. Now I use the pattern recognition tool for training. My question is how to use trained Neural Network in Matlab for classification. I've got the trained Neuronal Network and a picutre of a leaf, where each Pixel shall be classified as health or sick. Which function can i use ?

Answer 1

[Greg Heath]

Once you train an I-H-O neural net classifier net you can save it.

 save net1

At some future point when you want to classify M examples in the xnew matrix with

 size(xnew) = [ I M ],

Get the net from storage

 load net1

Then use the net:

 ynew = net1(xnew);

Hope this helps.

Thank you for formally acceting my answer

Greg

[Jan]

Thanks Greg it does work and excuse me for late response

Answer 2

[the cyclist]

I suggest you start here NNT help or here NNT video.

[Jan]

Thanks for help. I've already wachted the video and read the examples. But as far as I understood they use the pattern recognition tool just for training, because you have to committ the inputdata (attributes) and the outputdata (classes). But which function can I use (after training) to classify 'new' input data, in that case pixelinformation/wavelengths. Cause the aim is to classify each pixel.

[the cyclist]

On this page http://www.mathworks.com/help/nnet/gs/classify-patterns-with-a-neural-network.html, Step 5 under "using command-line functions" discusses actually using the neural network with new inputs.

Answer 3

[Greg Heath]

 ynew = net(xnew);

Hope this helps.

Thank you for formally accepting my answer.

Greg