Facial Recognition Fraught with Inaccuracies
November 2, 2016
Images of more than 117 million adult Americans are with law enforcement agencies, yet the rate of accurately identifying people accurately is minuscule.
A news report by The Register titled Meanwhile, in America: Half of adults’ faces are in police databases says:
One in four American law enforcement agencies across federal, state, and local levels use facial recognition technology, the study estimates. And now some US police departments have begun deploying real-time facial recognition systems.
Though facial recognition software vendors claim accuracy rates anywhere between 60 to 95 percent, statistics tell an entirely different story:
Of the FBI’s 36,420 searches of state license photo and mug shot databases, only 210 (0.6 per cent) yielded likely candidates for further investigations,” the study says. “Overall, 8,590 (4 per cent) of the FBI’s 214,920 searches yielded likely matches.
Some of the impediments for accuracy include low light conditions in which the images are captured, lower procession power or numerous simultaneous search requests and slow search algorithms. The report also reveals that human involvement also reduces the overall accuracy by more than 50 percent.
The report also touches a very pertinent point – privacy. Police departments and other law enforcement agencies are increasingly deploying real-time facial recognition. It not only is an invasion of privacy but the vulnerable networks can also be tapped into by non-state actors. Facial recognition should be used only in case of serious crimes, using it blatantly is an absolute no-no. It can be used in many ways for tracking people, even though they may not be criminals. Thus, it remains to be answered, who will watch the watchmen?