By Karl Ricanek Jr.,
NOTE: This is an overview of the entire article, which appeared in the September 2014 issue of Computer magazine.
Click here to read the entire article.
This article explores the use of biometrics, in particular soft biometrics — human physiological and behavioral characteristics that don’t explicitly identify individuals but can help differentiate them — for novel applications- through the lens of analytics.
Biometric analytics is the discovery of useful patterns within biometric signals to ascertain potentially interesting information about a person other than identity, such as emotional state or longevity. This emerging approach could one day dwarf the use of biometrics for identification in the commercial sphere given the public’s heightened sensitivity to such applications. Biometrics pioneer Joseph Atick acknowledged that the unfettered proliferation of biometric identification for consumer devices and apps (e.g. Google Glass and Facebook) is “basically robbing everyone of their anonymity.”
Biometric signals are the starting point for all biometric systems. Signals associated with common biometric characteristics include 2D ridge-valley maps in fingerprints, audio recordings of the voice, near-infrared images of the iris, and video of gait motion. Biometric algorithms extract various features relevant to identification. Researchers continue to develop new ways to harness such signals’ power – for example, using electrocardiography (ECG) waveforms to achieve aliveness detection and continuous authentication.
Researchers at the University of Southampton, UK, are exploring ways to improve recognition through gait analysis, which has thus far been largely limited to medical and sports science applications. They have also developed a technique that extracts soft biometric features of subjects captured in CCTV video including height, stature, position/pose, and clothing typed/color to help identify persons of interest.
In terms of commercial uses, several companies have developed biometric analytic systems that go beyond recognition to analyze voice recordings and face images. Voice recognition is used to better understand a speaker’s emotional state, honesty, concentration level, and other attributes to define a person’s character and personality. Face analytics software, developed by numerous start-ups, aims to “digitize emotion, so it can enrich our technology, for work, play and life.” Another unique use of face analytics is age estimation which will help in determining longevity and would also have implications in the cosmetics and rejuvenation industries.
What will this bring us in the future? Read the full article for more insight.
ABOUT THE AUTHOR
Karl Ricanek Jr. (firstname.lastname@example.org) Identity Sciences column editor, is director of the Face Aging Group at the University of North Carolina Wilmington.