Follow us on Twitter
Follow us on FacebookCCTV turns into Mystic Meg
@ 2012/10/29In a development which looks like it comes from a bad episode of Red Dwarf, researchers from the Carnegie Mellon university researchers have developed a surveillance system which spies on humans and can predict what might happen next.
According to Parity News the researchers have been developing the spying software for the Army. Dubbed Mind's Eye, it recognises human activities in video and can predict what might happen next.
This means that the coppers can be on the scene before anything nasty happens. Or people can be arrested before a crime takes place. If the software gets good, it can predict if a child will commit the crime at the moment of birth and make sure it pays for it, long before it even thinks of it.
Curiously this is exactly the software which was demonstrated in an episode of Red Dwarf, where such predictive software sabotaged a repair on the assumption that the human crew would stuff it up.
In a paper with the catchy title "Using Ontologies in a Cognitive-Grounded System: Automatic Action Recognition in Video Surveillance," the researchers describe an artificial intelligence cognitive engine that automatically detects and interprets a person's actions using a surveillance feed.
It is all based on previous research in the field of surveillance and specifically in object recognition.
They have also incorporated activity forecasting that tries to understand what humans will do based on their physical trajectories.
Currently the main cost associated with CCTV surveillance are the people needed to analyse the imagery.
If this sort of software works then local councils will be tempted to use it to replace humans involved in imagery analysis. What could possibly go wrong?
According to Parity News the researchers have been developing the spying software for the Army. Dubbed Mind's Eye, it recognises human activities in video and can predict what might happen next.
This means that the coppers can be on the scene before anything nasty happens. Or people can be arrested before a crime takes place. If the software gets good, it can predict if a child will commit the crime at the moment of birth and make sure it pays for it, long before it even thinks of it.
Curiously this is exactly the software which was demonstrated in an episode of Red Dwarf, where such predictive software sabotaged a repair on the assumption that the human crew would stuff it up.
In a paper with the catchy title "Using Ontologies in a Cognitive-Grounded System: Automatic Action Recognition in Video Surveillance," the researchers describe an artificial intelligence cognitive engine that automatically detects and interprets a person's actions using a surveillance feed.
It is all based on previous research in the field of surveillance and specifically in object recognition.
They have also incorporated activity forecasting that tries to understand what humans will do based on their physical trajectories.
Currently the main cost associated with CCTV surveillance are the people needed to analyse the imagery.
If this sort of software works then local councils will be tempted to use it to replace humans involved in imagery analysis. What could possibly go wrong?
