Yes for the face they measure your iris and sclera and the distance between your eyes. They compute a ratio output and use that. If your nose and mouth are visible they also create ratios for those. Distance and hight x length . By modifying these points, the ratios are poisoned.
For movement gait identification. They use stride and leg lengths. (Thighs and calf ratios). Speed. Overall height. And where legs start (,hip). They make ratio measurements. But that’s not as useful as it is for face. Not as accurate because of clothing.
They use something more like motion capture.
You ever seen videos of people in black skin suits with ping pong balls glued to it. ? They use this in games and cgi.
They track movement with the ping pong balls and then map it on a computer to the CGI model.
This is what is basically used for gait. tracking.
But. Maybe you are thinking. “Holdon. If they need to use the ping pong balls in studios to capture motion, how can they do it from a video of someone walking down the street?”
Well. Firstly it’s not nearly as good as the suit/ball method done in a studio.
But it doesn’t have to be to create a high enough resolution of the movements to create a signature .
That said. Unlike facial recognition, it’s current ability to identify someone is less accurate and they can’t just run it through a big database. Instead they use other information like cell phone pings to first create a list of possible identities.
As you can guess. That still means there is huge amounts of error.
There are huge amounts of error in facial recognition too. They just downplay it.
But it’s not nearly as good as they want you and juries/judges to think.
Anywho. This is why the dazzle camouflage is specifically useful for gait recognition poisoning.
The multiple high contrast bars of black and white confuse the software . The software cannot find the actual outline of the person vs the background. Especially in a video from afar. It poisons the data.
Even modern day equipment has low accuracy determining the distance and specific location of things with this camouflage.
What exactly are the cameras scanning for? Is it just measurements between specific points? It would probably work pretty well for that.
Yes for the face they measure your iris and sclera and the distance between your eyes. They compute a ratio output and use that. If your nose and mouth are visible they also create ratios for those. Distance and hight x length . By modifying these points, the ratios are poisoned.
For movement gait identification. They use stride and leg lengths. (Thighs and calf ratios). Speed. Overall height. And where legs start (,hip). They make ratio measurements. But that’s not as useful as it is for face. Not as accurate because of clothing.
They use something more like motion capture. You ever seen videos of people in black skin suits with ping pong balls glued to it. ? They use this in games and cgi. They track movement with the ping pong balls and then map it on a computer to the CGI model.
This is what is basically used for gait. tracking.
But. Maybe you are thinking. “Holdon. If they need to use the ping pong balls in studios to capture motion, how can they do it from a video of someone walking down the street?”
Well. Firstly it’s not nearly as good as the suit/ball method done in a studio.
But it doesn’t have to be to create a high enough resolution of the movements to create a signature .
That said. Unlike facial recognition, it’s current ability to identify someone is less accurate and they can’t just run it through a big database. Instead they use other information like cell phone pings to first create a list of possible identities.
As you can guess. That still means there is huge amounts of error.
There are huge amounts of error in facial recognition too. They just downplay it.
But it’s not nearly as good as they want you and juries/judges to think.
Anywho. This is why the dazzle camouflage is specifically useful for gait recognition poisoning.
The multiple high contrast bars of black and white confuse the software . The software cannot find the actual outline of the person vs the background. Especially in a video from afar. It poisons the data.
Even modern day equipment has low accuracy determining the distance and specific location of things with this camouflage.
First-rate comment! Thanks!