Quote:
Originally Posted by tucker6
would you bet your life on it?
|
Funny you should ask that.
Is that technology sufficient at the present time?
In most cases I'd tend to say yes.
But in some cases my answer would be hell no. At least not until such time as the technology is vastly improved.
Using technology that is currently available as I type this:
On a clear sunny day - with no obstructions whatsoever between cameras and roadside signs - computers programmed with TDA based image recognition software should be able to very quickly identify objects such as stop signs, yield signs, or speed limit signs with near 100% accuracy.
The software is able to do that based on shapes and colors. Or more specifically after having analyzed the coloring and locations of pixels in the tens of thousands or hundreds of thousands of images it has been fed during training sessions.
In the case of a stop sign the software should be able to discern the octagonal shape and the red coloring into meaning: That's a stop sign.
In the case of a speed limit sign the software should be able to recognize the shapes and coloring of the individual numbers and letters on the sign and translate those into a speed limit.
But what if some idiot does something stupid?
For example - hangs a beach towel over top of a stop sign?
Suddenly the shape is only octagonal on one side - with something the software hasn't been programmed to recognize flapping around in the breeze on the other side.
It's that "what if some idiot does something stupid" category that concerns me.
Most of us could probably think up a 1000 different ways someone could do something stupid to trip up the system.
-jp
.
