3 ms·

Doesn't look like it totally worked, based on that last image. Pretty good though. I wonder if it'll be possible to come up with something that generalizes to a

• pelorat 4y ago
  No, that's not possible. This approach can only fool models that identifies people as a whole based on a 2D array of pixels. It can't fool models that understand that a person is a composite of other features. If you create models that can detect individual features like eyes, nose, ears, head, arms, hands, legs, neck, fingers, hair and further feed those into other models that can validate how parts belong together for them to form a human (pose estimation, outline detection to 3D mesh mapping), this research would not work at all. This research works only on straightforward single network models where input is a 2D array of pixels, and output is a yes/no.