surreptitious57 wrote: ↑Sat Feb 09, 2019 8:01 pm
X rays which have been around for a hundred years can easily distinguish between human and non human bodies
What?
No they don't. All that x-rays do is produce images.
From the image you still need to determine human (1) or not-human (0).
You can use computer vision here:
https://en.wikipedia.org/wiki/Computer_vision
surreptitious57 wrote: ↑Sat Feb 09, 2019 8:01 pm
There is also DNA which can be used to identify any body that is not mono zygotic which is virtually all of them
These are not going to be problems for the machines of the future that will replace us as the dominant species
Our reptilian / mammalian brains will be history by then and the human race may even be on the slow path to extinction
That's a far more viable strategy. At the median it will probably make very few errors, but eventually you are going to run into the exact same problem as cutting up the color spectrum into boxes.
Because "human" is a made up category so that we can be inclusive of everybody and grant them "human" rights you will actually find that the genetic distance between cultures who haven't been in contact for generations (say Danes and Australian Aboriginees) is drifting very far apart.
But, as we know - there are no perfect binary classifiers and inevitably you are going to bump into problems. False positives and true negatives...
https://www.medcalc.org/manual/roc-curves.php
You inevitably run into the problems of sensitivity and specificity.