I have a rudimentary XOR trained neural network working correctly with the following structure. 2 Inputs, 2 hidden nodes and 1 output. I would like to extend this to grayscale image recognition with NxN inputs, M hidden nodes and O outputs.
My question is, does/would the same underlying theory hold true for a network with NxN floating point input values between 0.0 and 1.0 used in the same feedforward and backpropagated application as the one used for the XOR network, or are there more steps involved?
Assuming that you want to create a classical network - everything is exactly the same, simply vectorize your images - meaning that it is not really NxN matrix, but rather N^2 long vector of values between 0 and 1.