How to find anyone anywhere with online facial recognition
Facial recognition technology is turning up in ever more applications – from the useful, like unlocking smartphones, and the fun, like Facebook tagging, to the essential, like crime detection, or the life-saving, like prevention of terrorism.
Our faces too are photographed, filmed and sometimes clocked almost everywhere we go. We post them ourselves, on social media or elsewhere on the web. How many images of your face does your name yield in Google Images? Mine turns up a few dozen, half of them appearing at the top of this monthly column. I’ve hardly aged! It’s not as many as the Queen or David Beckham, but then I manage to avoid the paparazzi and rarely post selfies (I made an exception when I met Giorgio Moroder at CES in January).
I don’t really want my phizog everywhere online, for no particular reason really except vague, probably irrational worries about security and privacy. I would be more concerned if I lived in a more repressive regime, especially if was looking to change it.
So how easy is it to search the internet and find anyone anywhere? Policies on facial recognition vary widely around the world; some governments employ it freely themselves while others are more cautious about citizen privacy. Some allow private companies more free rein than others. And the tech giants can themselves also be cautious about the implications of allowing anyone to find any face anywhere on the web. Upload a picture of yourself to Google Images and it will produce people with similar clothing or backgrounds but probably not you.
Yet it may only be a matter of time before the genie is really out of the bottle because it is so very easy. Facial-recognition technology is freely available as open-source code packages. Ben Heubl tried it for E&T. It was scarily or satisfyingly efficient, depending on your viewpoint. We also tried it with a picture of Lord Lucan to see if we could solve that mystery. Most of the matches were taken before his disappearance, but it also flagged up James Coburn in ‘A Fistful of Dynamite’ as a match. Who would have guessed? Yes, it works but it’s not perfect.
Also in this issue, we start a new regular feature with TV presenter Dr Shini Somara interviewing some extraordinary engineers about their careers, influences and aspirations. First we hear from Clare Elwell, who develops optical monitoring and imaging systems for medicine at UCL, about what makes her tick.
Next month we’ll be revealing the shortlists for the E&T Innovation Awards. Fingers crossed if you’ve entered, and if you haven’t – well, do it next year! It will be a virtual event this year, but there are some exciting ideas in the pipeline and we aim to make it bigger and better than the real live event. Stay tuned and, as the autumn approaches, stay safe.