Facial recognition technology matches human faces from digital images or videos to those stored in a database of ‘known’ faces.
It is commonly found in authentication systems — for instance, in some models of Apple’s iPhone. But the technology can also be found in surveillance systems.
What is it?
Broadly speaking, facial recognition is a system of matching a human face captured in a digital image or video stream to database records.
These systems have existed for decades but recent improvements in pattern recognition have turbocharged interest in facial recognition systems.
Typically, facial recognition is treated as a method for authentication. That said, there is growing interest in some subsets of facial recognition, such as expression recognition — where the aim is to recognize whether a subject’s emotional state. This technology is being used by some retailers to gauge customer interest in their products.
What’s in for you?
As an authentication system, facial recognition is convenient and touchless — which may appeal to some.
And some companies — for instance retailers and airports — have trialled the technology for security purposes.
What are the trade offs?
While the systems can be highly effective in perfect conditions, they tend to be less reliable when in real-world conditions, with crowded environments, variable lighting and often less-than-ideal camera angles.
More worryingly, many facial recognition packages are poor at correctly identifying faces of anyone other than white males. Given they’re often used in surveilling crowds, it’s not surprising that many people have concerns over the use of the technology.
The technology is also readily defeated. People not wishing to be identified can wear face masks or paints.