There are also things like 123D Catch, a free online service, that will take a set of photographs of an object, taken from different positions, and turn them into a 3D model that can be given to a CAM programme to produce gcode. It does work, I have used it, but it doesn't work anything like as well or as easily as its documentation implies. Took me days to edit the STL 3D model data to get something usable. But in the end it did work. It's an option, and it might be better now than when I last used it (around a year ago).