Interesting... the preset menus don't make any sense (they have different headers, but just repeat variations on the same 20 or so presents) and most of the descriptions for those presets are only vaguely related to the images they create. How much of the design, programming, and bug testing for this did you leave to a Neural Net Model?