Re: Spooking of neural nets and image recognition...

As to serveilence cameras, is anyone familiar with methods of not getting
photographed in the first place? Some methods I've heard of:

1) spray paint or vasoline on the lens (kinda obvious :)

2) putting a pin or finnishing nail in it's coaxial cable (umbilical if you
  will) - how good of a job of scrabbling the image will this do?

3) mounting a "test pattern" in front of the lens.

4) placing a magnet on the side of the camera (does this work?)

5) timing the pan, and taking advantage of "dead areas" (areas not covered
  for brief periods). Anyone familiar with the general range/resolution
  ratio, and how wide an area will get covered?

  Any other methods?

How about radio-frequency interference?   I've heard that this works
quite well for video camera and other electronic monitoring equipment.

Ben Byer <bbyer@bix.com>