The "10-year-challenge" posts currently trending on social media are ideal learning materials to teach machines to help catch criminals a Wired columnist believes – and those posts could play a role in helping SA find missing children in the future.
The posts, trending on Facebook and Twitter since the beginning of January, asks users to share photos of themselves from 2008, and then again in 2018.
In a opinion piece published on Wednesday, Wired’s Kate O’Neill explained how the posts could help improve algorithms designed to take into account ageing.
Highly accurate facial recognition systems already exist, and the technology is already in everyday use, such as for unlocking your phone just by showing it your face, but they rarely take into account ageing.
The 10-year-challenge provides a huge amount of data easily identified with keywords such as “me in 2008” and “me in 2018” O’Neill said .
This in contrast to typical Facebook photos, which may not be posted chronologically, or profile pictures that often don't feature the user themselves.
10-year-challenge information, and systems developed using it, could help law enforcement agencies across the world identify criminals or missing people who’ve aged, O'Neill said.
Facial recognition is already widely used to catch criminals around the world.
In 2017, Chinese authorities used facial recognition to identify 25 suspects who had been avoiding arrest, including a man who had been on the run for a quarter of a century.
The Guardian reported that Qingdao authorities said their system had a 98.1% accuracy rate.
A year later, a man was arrested in China’s Jiangxi province after being picked out of a crowd 60,000 after he had been on the run for “economic crimes”, the Washington Post reported.
In India, authorities used facial recognition to find 3,000 children within four days in April, 2018.
Facial recognition has also been used by United States immigration officials to identify illegal immigrants.
Also from Business Insider South Africa: