This month, Computing Now explores biometrics—the authentication of persons based on physical or behavioral characteristics. Biometric technology can be found in many aspects of our daily life, from paying for groceries to accessing personal computers and buildings to automatically labeling and organizing digital pictures.
Biometrics has a long history dating back several thousand years, when biometric information was used to process payment (food rations) for thousands of workers building the pyramids. Historical documents indicate that notes were kept detailing names and birthplaces along with physical characteristics (height, weight/girth, scars, deformities) and behavioral traits (rude, sullen, braggadocio, happy) to ensure that workers did not cheat the payment system.
In more modern times, biometrics can be traced back to work by Henry Fauld (1880), Francis Galton (1888), and Edward Henry (1899) on identification based on fingerprint patterns. The events of 9/11 instigated more recent research and development activities in biometrics. The premise of this research is to discover techniques that could uniquely distinguish a potential terrorist from the general population (1-to-N-million search). The outgrowth of this effort has been efforts to develop biometric systems that can verify identity (verification), which performs 1-to-1 matches, a much easier task than identification.
The fevered pace of biometrics research has created new modalities based on keyboarding patterns or mouse movements, walking patterns (gait), types of utterances (speech), the configuration of veins in the finger or hand (veinal), geometries of the finger or hand, the face, and the complex structures of the melanin-rich area of the eye (iris). These emerging biometric modalities have created vast commercial opportunities outside the more common public sector uses. In fact, Acuity Market Intelligence, a technology strategy company located in Louisville, Colorado, predicts a compound annual growth rate of 19.69 percent for biometric technology from 2009 to 2017, at which point commercial deployment of biometrics will outpace public sector use.
This month’s articles showcase the diversity of biometric technology and its uses by describing popular biometric modalities including fingerprint, face, and iris recognition; discussing what’s next for these traditional techniques; and highlighting new modalities that may soon be available as commercial products.
Four foundational articles are taken from Computer’s February 2010 special issue on biometrics.
In “Iris Recognition: The Path Forward,” (login required for full text) Arun Ross provides a systematic overview of the components of an iris recognition system, describes its prowess for identification, and indicates directions for improvements and future research.
In “Fingerprint Matching,” (login required for full text) Anil K. Jain and coauthors provide a thorough review of the use of fingerprints, which are unique to each person, from the inception of this technology in the mid-1800s to law-enforcement and civilian applications today. The authors point out that, although fingerprints have been used for identification for more than a century, challenging research problems remain in perfecting automatic recognition.
“Face Recognition by Computers and Humans” (login required for full text) by Rama Chellapa and colleagues details the difficulties with automatic face recognition by computers, discusses the mechanics of human face recognition, and compares the two approaches.
The final article from Computer’s special issue, “Unconstrained Biometric Identification: Emerging Technologies,” (login required for full text) discusses the next generation in biometrics for identification, which removes the current constraints placed on face and iris recognition. The authors articulate the face recognition problems associated with facial aging and introduce soft biometrics in the form of age estimation from a face. The remaining articles discuss various uses of biometrics from streamlining border crossings to the intricacies of matching a person to new modalities for verification.
In “Biometrics Could Streamline Border Crossings,” (login required for full text) Greg Goth provides compelling arguments for the use of biometrics in travel security and discusses the necessity for governments to work together to define the use of the technology.
“The Biometric Menagerie,” (login required for full text) by Neil Yager and Ted Dunstone, offers insights into why it’s more difficult to authenticate some people using biometric systems. This article distinguishes matching problems associated with “goats” (subjects that are difficult to match), “wolves” (subjects that match well to others), and “doves” (subjects that match well only to themselves).
“A Novel Approach to Design of User Re-Authentication Systems” (login required for full text) by Harini Jagadeesan and Michael Hsiao, and “Biometric Authentication Based on Infrared Thermal Hand Vein Patterns” (login required for full text) by Amioy Kumar and coauthors, introduce two little-known biometrics. The first article provides experimental data on the efficacy of re-authenticating a PC user based on mouse and typing gestures. The second article discusses evidence for a new physical biometric system based on hand veins. The system acquires the biometric signal from infrared images of a hand, which are computed by Gabor filters to produce a biometric template for matching.
We hope you find these articles on biometrics informative and enjoy reading them as much as we have.
Ron Vetter, the cofounder of Mobile Education, LLC, and a professor in the Computer Science Department at the University of North Carolina, Wilmington, is a member of Computer’s editorial board. Contact him at vetterr at uncw dot edu.
Karl Ricanek Jr. is the director of the Face Aging Group Research Lab and an associate professor in the Computer Science Department at UNC Wilmington. He has worked in the area of biometrics since 1997. Contact him at ricanekk at uncw dot edu.