This market research report was originally published at Tractica's website. It is reprinted here with the permission of Tractica.
Facial recognition is a popular application of computer vision technology that deals with recognizing the identity of a person. The technology has been in use for several years with varying degrees of accuracy. Classic facial recognition techniques, such as scale-invariant feature transform (SIFT) and speeded up robust features (SURF), relied on extracting the unique features of a face. These techniques compared values of the incoming picture with a reference picture to generate a match. While this worked fairly well in situations where the incoming picture featured similar conditions (e.g., a photograph with a similar background), it was generally inadequate for detecting a person’s aging or a partially covered face, so the accuracy remained limited. Despite its limitations, facial recognition was used frequently for security and surveillance, and law enforcement agencies have used it for quite some time.
Given the advances of artificial intelligence (AI), new techniques for facial recognition have emerged. The accuracy of results has increased drastically. The technology has advanced beyond recognizing facial features to detect emotions and even diseases. Historically, a personal computer (PC)-class device was required to run such algorithms, but the falling prices of hardware enables running facial recognition on a camera, making it easy to employ new capabilities in different places. This is giving rise to a wide range of new applications via the facial recognition technology.
Real World Facial Recognition Technology Applications
Apple’s latest iPhone is the most prominent application of late for general users. The technology allows users to unlock the device by simply looking at the screen. Recently, Apple also announced the availability of face detection application programming interfaces (APIs) that developers can embed in their applications. All of this runs on the device itself, making it easy to create edge applications.
The Southern California-based CaliBurger chain is taking facial recognition into the retail space. It is testing a facial recognition-based system for ordering burgers. The system links a loyalty customer’s face to their past orders. For subsequent orders, the customer just needs to look into the camera and their orders will automatically pop up.
In Boston, the Department of Homeland Security (DHS) is testing a system that lets travelers check in via facial recognition rather than having to show a boarding pass. For certain international flights in Atlanta and New York, the DHS has partnered with Delta to bring mandatory facial recognition scans to the boarding gate. The Delta system checks that a passenger is supposed to be on the plane by comparing the facial image, which is captured by a kiosk at the boarding gate, to passenger manifest photos from the U.S. Department of State’s databases.
Walmart is testing a facial recognition system to identify unhappy customers. The technology uses video cameras at store checkout lines that monitor customers’ facial expressions and movements. If the behavior is deemed to be “unhappy,” the system alerts its employees.
Facebook has also recently launched a new facial recognition feature called Photo Review. The feature alerts users when their face shows up in newly posted photos. The user has the option to tag themselves, leave as is, or ask the uploader to take the photo down, or even report it.
Reactions to and Opportunities Presented by Facial Recognition Systems
The protocol in the case of a failure in any of these systems is still unclear. If the burger chain’s system that recognizes a face fails, it probably will not cause any problems. However, if a person is denied boarding on a flight due to a failure of the DHS system, it is sure to generate a lot of unwanted media attention.
Facial recognition technology is evolving at a rapid pace, and more and more places will soon know their customers and users by their facial images. It is unclear how this will impact human behavior and if there will be any social pushback to such a system. One thing that is certain is that when it is used wisely, facial recognition can open up many new business opportunities.
By Anand Joshi
Principal Analyst, Tractica
