Facial recognition software has come a long way in the past few years. While early face recognition software needed human input at every step, it’s advanced enough today that it can compare a single face against millions of faces on record in just a few seconds.
Thanks to facial recognition software like Face Unlock, the private date on your cell phone can be kept secure and accessible only to you. Criminals and terrorists on the run can be caught in record time whenever their faces appear on camera.
Some applications still have a long way to go, though. For instance, Face Unlock is not as secure as other phone locking methods, and the FBI has trouble matching faces in images that aren’t well-lit or looking at a face straight-on.
Still, the development of facial recognition software is moving forward in leaps and bounds. Companies are coming up with more and more uses for the software, from dating websites matching users based on the looks of the exes, to advertisers displaying targeted ads based on how facial recognition software determines your age and gender.
But developing just as rapidly as facial recognition technology is the fear and suspicion of exactly how the new software will be used. With government violations of privacy all over the news today, it’s easy to understand the concerns.
Who exactly is recording your facial measurements and identity? Where are they tracking you, and how are they using all that data? Can we really trust those in power to use our data responsibly, and keep it secure from those who would abuse it? How can we protect our faces from ever-present cameras and scans?
Companies like CV Dazzle are discovering ways to fight back against facial recognition with custom hairstyles and makeup that baffle the software. But there may not be a guaranteed way to escape it completely. As long as facial recognition software is being developed and perfected, it’s important to keep informed about who is compiling your data and how they’re using it.
Transcript: Face Recognition
Facial recognition technology: whether it unlocks your smartphone or spots a potential terrorist in a crowd, it seems like a smart solution to a range of problems.
However, in an age of civil liberty scandals, how much are we really being watched?
What’s Facial Recognition?
Facial recognition technology uses images of individuals captured from video cameras, websites, and pre-existing photographs.
How It Works
- A still image of the subject’s face is analyzed for distances between eyes, nose, mouth and jaw.
- These distances are converted into a digital algorithm.
- The algorithm is compared to a database of previously analyzed faces until a match is made.
Who’s Watching You Today?
55,000 – The number of “facial recognition quality images” intercepted by the NSA daily.
Documents obtained by Edward Snowden reveal the extent of the spy agency’s efforts to monitor individuals both on and offline.
$5,000 – Price of a 6-month premium online dating membership, complete with personal matchmaker.
Companies such as Match.com and Three Day Rule offer premium services that use facial recognition software to pair up lonely hearts based on pictures of their exes.
1.8 million – The number of Yahoo user accounts that Britain’s surveillance agency, GCHQ, collected webcam imagery from globally, over a 6-month period.
GCHQ admitted that “a surprising number of people use webcam conversations to show intimate parts of their body to the other person.”
90% – Success rate of the OptimEyes system in telling males and females apart
Supermarket checkouts are being equipped with face detection software that can recognize a person’s gender and age group and run tailored advertisements.
10,000+ – The number of public and private cameras accessible by Chicago authorities
London is likely the most observed city in the world, with more than 91,000 cameras in place throughout London boroughs and the city’s transit network.
Who’s Watching You Tomorrow?
- Tagging suggestions could become much more accurate:
- 97.53% – Humans
- 97.25% – Facebook’s DeepFace software
- Success rate for identifying two different photos are of the same unknown person’s face
- DeepFace is still in development, but early results show near-human levels of identification.
- US Patent No. 8,600,120 B2
- “Personal computing device control using face detection and recognition.”
- Whether Apple’s patented technology will be reserved for biometrically unlocking your iPhone or more is yet to be seen.
- Google has banned facial recognition apps for Google Glass, stating:
- “We won’t add facial recognition features to our products without having strong privacy protections in place.”
- However, users can still load facial recognition apps on to Google Glass without Google’s permission.
The Future of Facial Recognition
3D facial recognition will improve success rates, capturing the contours of eye sockets, nose and chin.
Skin texture analysis will convert the unique lines and spots on a person’s skin into a mathematical space – improving identification by 20-25%.
- Developed by Brooklyn-based artist Adam Harvey, CV Dazzle is a project that uses state-of-the-art hairstyles and cosmetics to fool facial recognition algorithms.
As technology improves, our privacy will be even more at stake. But are we heading towards a safer society or an Orwellian dystopia?
- Facebook Creates Software That Matches Faces Almost as Well as You Do – technologyreview.com
- with Facial Recognition Partnership, Match.com Users Can Find Dates Based on Exes – mashable.com
- about Three Day Rule – threedayrule.com
- Looking for Help and Advice? – match.com
- DeepFace: Closing the Gap to Human-Level Performance in Face Verification – facebook.com
- Face Recognition – hrsid.com
- Personal Computing Device Control Using Face Detection and Recognition – uspto.gov
- Smartphones and Facial Recognition: Focus Groups 2.0 – theguardian.com
- Facial Recognition: Is the Technology Taking Away Your Identity? – theguardian.com
- Google ‘Bans’ Facial Recognition on Google Glass, but Developers Persist – theguardian.com
- Quividi Defends Tesco Face Scanners after Claims over Customers’ Privacy – theguardian.com
- Tesco’s Plan to Tailor Adverts via Facial Recognition Stokes Privacy Fears – theguardian.com
- Match.Com Uses Facial Recognition Software to Pair Users with People Who Look like Their Exes – onthemedia.org
- Facebook’s DeepFace Software Can Match Faces with 97.25% Accuracy – forbes.com
- NSA Collecting Millions of Faces from Web Images – nytimes.com
- Optic Nerve: Millions of Yahoo Webcam Images Intercepted by GCHQ – theguardian.com
- Facial Recognition Fails in Boston – businessinsider.com
- Embracing Big Brother: How Facial Recognition Could Help Fight Crime – edition.cnn.com
- CV Dazzle – cvdazzle.com
- How Facial Recognition Works, Face Recognition, Facial Biometric Technology – Ex-sight.com
- Facial Recognition – findbiometrics.com
- Face Recognition Study: FAQ – heinz.cmu.edu
- Walter Mitty’s eHarmony Counselor Costs $5,000 – marketwatch.com
- ACLU Report: Chicago’s Video Surveillance Cameras: A Pervasive and Poorly Regulated Threat to Our Privacy – scholarlycommons.law.northwestern.edu
- Surveillance Cameras Sought by Cities after Boston Bombs – bloomberg.com