straightpick wrote:Well, first of all, a macro (closeup) of a key and a shot from 150 feet away are not going to reveal the same information. Ever work with multiplex keyways? The differences in the millings are so minute I have trouble distinguishing them when I have the blanks in my hand. Plus there are two sides of a key. You will probably only see one side with the camera, which will leave a lot of guesswork as to what the other side might be.
Don't get me wrong, I definitely agree with the points you make here. I think the most important thing to consider, however, is that most privilege escalation and unauthorized access non-destructive attacks don't come from total strangers; they come from people already holding valid keys. It's absolutely trivial these days for someone who owns their neuter-bow key to take a couple sufficient quality closeups with a digital camera, which is all that you should need for a computer vision system to identify the blank type and the bitting.