There are few pieces of technology used by security companies that are more widely discussed than facial recognition systems.
A form of biometric security akin to voice recognition, fingerprint or retina scanning, facial recognition is a collection of different technologies that matches the faces of people who walk past certain suitable cameras to a database of faces.
This is often used in event security to identify people who are on a watchlist as a person of interest, although it is typically used as the start of an investigation in combination with other evidence rather than as definitive proof in and of itself.
It is a constantly evolving technology due to its complexity, and the greatest proof of how difficult it has been to use can be found in the fact that the technology is over six decades old and is still fundamentally changing.
The first pioneering experiments in facial recognition technology were attempted by artificial intelligence pioneers Helen Chan Wolf, Woody Bledsoe and Charles Bisson.
The technology was exceedingly primitive, even requiring a human to point out where particular facial features were located such as the centre of the pupils, the widow’s peak and the corners of each eye.
Once these details were added, the system could be given an image, compare the differences in features and find potential matches.
The first automatic facial recognition system was demonstrated in 1970 by Takeo Kanade, one of the foremost researchers in the field of computer vision, or the discipline of developing AI systems that can scan photographs the same way human eyes can.
The system could “look” at a photograph, locate facial features and calculate where facial features are without a human wielding a graphic tablet. It ultimately was not terribly reliable, but it proved the principle over two decades before the first practical systems began to be developed and used.