Skip to main content

The Promise & Pitfalls of Facial Recognition & Biometrics

The Promise & Pitfalls of Facial Recognition & Biometrics
  • Those of us involved in the pioneering days of digital signage were introduced to personalized programmatic advertising via biometrics through Steven Spielberg’s 2002 film Minority Report. The scene opens as the film’s main character, John Anderton, played by Tom Cruise, is fleeing the authorities and traveling through a future shopping mall environment. As he moves through the environment he is continually ‘tracked’ by way of retinal scanners and personalized advertisements from Lexus, American Express and The Gap, all of which present themselves on ‘digital signs.’ One display addresses him directly by name - “John Anderton… you could use a Guinness right about now.” Not unlike George Orwell and Gene Roddenberry, Philip K. Dick and Steven Spielberg foreshadowed a world in which brands use biometrics to serve personalized advertisements. Almost 15 years later this one form of biometric technology- facial recognition is no longer science fiction and is rapidly becoming commonplace.

While this scene foreshadowed the promise and potential of digital signage, it also portended the nefarious use of technology to invade our privacy as foretold by George Orwell’s ‘big brother.’ In today’s world, cameras surround us from surveillance, to our laptops to our mobile phones. Google, Facebook, Apple and Microsoft all leverage facial recognition technology in their products and services. While Intel may have coined the phrase Anonymous Video Analytics (AVA) with their acquisition of Cognovision in 2010, being anonymous in public spaces is very likely a thing of the past. Whether it’s a surveillance camera, your credit card, your car, your webcam, your thermostat, your baby monitor or your mobile device, you ARE being monitored and tracked. This is simply a byproduct of ‘being digital.’

While there are no federal laws that specifically govern the use of facial recognition technology, both Texas and Illinois have laws against using such technology to identify people without their informed consent. That translates to one out of every eight Americans having a legal right to biometric privacy! A lawsuit filed in Illinois alleges Facebook violates the state’s Biometric Information Privacy Law (740 ILCS 14) by taking user’s faceprints “without informing its users- let alone obtaining their informed written consent.” Facebook’s algorithm, DeepFace, has an accuracy rating of more than 97 percent and is active by default. Google’s FaceNet algorithm is over 99% accurate, but is activated only if you opt in. Allegedly Facebook has already spent more than $10million dollars lobbying against this one state law.

While facial recognition captures an image and processes its characteristics and geometry in great detail for identification and comparison, select anonymous video analytics algorithms identify and recognize a human face by way of simple geometry that define two eyes and a mouth. This technology does not require storing and processing raw video footage or photographic images as part of its processing - hence the use of ‘anonymous’ in its definition. This is technology is often referred to as eye- or gaze-tracking software. In recent years, this technology and its algorithms have been enhanced to include demographic and emotional insights.

It is noteworthy that although the storage of images and/or raw video is not required for this biometric technology, the technology itself does not preclude its user from recording images and/or video. It is the client’s responsibility to clearly and contractually identify, define and monitor the ‘rules of engagement’ and use of this technology. Commercial examples of this technology include Nomi, Cenique, Kairos/IMRSV and Intel AIM but there are many open-source algorithms available for those who are interested in building their own mousetrap.

While privacy issues persist and are subject to differing philosophical perspectives, its safe to say that while the use of this technology has value it can also be used in ways that cross the line to being ‘creepy.’ I don’t have an absolute definition of this but let common sense be your guide. Always use technology respectfully and bring value to the audience. My experience suggests that if you’re simply using the technology just to justify and/or monetize digital signage content then it’s most likely not a sustainable business model.

Another useful resource to help guide the use, and recommended best practices, of facial recognition technologies is the Federal Trade Commission’s March 2012 report, “Protecting Consumer Privacy in an Era of Rapid Change.” The recommendations and guidelines put forth in this report are based upon three core principles. These include:

1.Privacy by design - Companies should build in privacy at every stage of their product and/or service design.
2.Simplified Consumer Choice - For practices that are not consistent or typical with the context of a transaction or a consumer’s relationship with a business, companies should provide consumers with choices at a relevant time and context.
3.Transparency - Companies should make information collection and use practices transparent.

Other relevant material and guidelines can be found in the World Privacy Forum’s Digital Signage Privacy Principles and POPAI’s Recommended Code of Conduct for Consumer Tracking Research.

I founded my consulting practice, StoreStream Metrics on the basis of providing in-store media-based shopper experience concepts and designs founded upon quantitative measurement and evidence. As such, StoreStream Metrics has pioneered the use of several different video analytics technologies. Valued metrics include viewing attention time, shopper traffic counting and directional behavior, dwell time (time spent by a viewer within an area of interest) and the calculation of the viewing behavior of a unique piece of content (‘clip’) used in a published playlist or rotation across an entire network.

In the final analysis, however, it is not these metrics that have value to the client. They are interesting but ultimately not truly valued. The real value to these biometric measurements is the role they play in the full shopper journey. They provide quantitative insights as to the size of the potential viewing audience of your digital signage content and then, if processed properly, an insight as to the conversion journey of that potential audience to viewers. The real value perceived by the client is answering the question, “Did I succeed and/or optimize my content in conveying my intended message to the maximum number of possible people?” In any retail or commercial environment this typically translates to “Did the use of digital signage help me sell more stuff?”

Author Adrian Weidmann will lead the webinar discussion entitled, “The Promise & Pitfalls of Facial Recognition & Biometrics” when the presents its 30-minute webinar on Thursday, December 1 from 2:00-2:30pm EDT. Adrian is a member of the Digital Signage Expo Advisory Board, which addressed the question of facial recognition and biometrics last August. Other answers can be viewed at