NEC announced a system that identifies people even if they are with their backs to the camera
The company NEC announced its “People Re-identification Technology”, a system that applied to video surveillance is able to know who is who in an image even if our face is hidden. What can go wrong?
Video surveillance is gradually fine-tuning their skills to have us all well controlled. Now, the Japanese company NEC announces a system that allows people to be identified by comparing images of the entire body previously taken by other cameras, so that it would not be essential to have an image of the face of the subject in question to obtain its identification in a particular frame.
With the suggestive name of “People Reidentification Technology “, the system compares the complete photographic images of a person – size, body shape, clothing – to look for coincidences with other images and determine the identity of the subject.
This way, NEC ensures that it is able to reveal someone’s identity even if the image is taken from behind or from the side. It also allows, according to what the company said, to recognize someone in places where there are many people or spaces with visual obstructions that prevent the face or body of a person from being completely seen.
For this, this multinational says that it has used “image analysis and deep learning technologies” combined with facial recognition technology, so it has achieved “high precision” results that are not based solely on facial images
Deep learning techniques have been used, says the company, to match the image of a person by automatically selecting parts of their image that the camera does register. “The rate of coincidence for a person can reach 90%,” says NEC.
The biometric identification, based on patterns obtained through the analysis of large amounts of data, entails increasing risks for the security and privacy of citizens. False positives, abundant in recent times, require human verification to offer a minimum of guarantees.
Advocacy groups such as ACLU have been warning for years that biometric identification through cameras, specifically facial recognition, involves some dangers such as the indiscriminate surveillance of citizens.
For its part, Big Brother Watch also fights against the proliferation of video surveillance cameras with automatic recognition systems, so widespread in the United Kingdom, and warns that this activity in public spaces “threatens privacy, freedom of expression and the right to association”.