I have been trying to use the nntool in MATLAB to create a neural network to analyze images. I am able to resize the images down to about 5000 pixels grayscale. I set each pixel as an input, then has a hidden layer of 4-5 nodes, then have one linear output node.

The problem is that as soon as I try to run it using 'train' the program just freezes up. Is there a max number of input nodes I should be using in a neural network? If so are there any ideas how I can consolidate the inputs?