My main reason for asking this question is that I want to pull out digitial information from an image, and use that information to create a training set for a neural network. So far, I have pulled out the RGB values of each individual pixel in a 3 channel image. This is the only usable data that I see that can be used as inputs to the neural network.
Does anyone have any other ideas as to what I could use from a regular image that could be truly represented 'digitally'. For example, could I do something along the lines of representing edges as simple floating point numbers that I could simply plug into a neural network?