AsianScientist (June 1, 2015) – Iftah yā simsim! (pronounced “if-toe-yay seem-seem”) Perhaps better known as “open sesame” from the folk tale collection One Thousand and One Nights, this Arabic phrase controlled access to a treasure cave owned by a group of thieves, but was accidentally discovered by the woodcutter Ali Baba.
Like the thieves who owned the treasure in the Ali Baba story, mankind has long sought effective ways to authenticate individual persons—be it for access control to ensure confidentiality, forensic purposes or criminal identification.
Handprints were already used as criminal evidence in China during the Qin and Han dynasties and as early as the 1850s, people realized that fingerprints could be uniquely determined by ridge characteristics and specific localized features known as minutiae. Apart from handprints and fingerprints, scientists have discovered that other biological, physical and behavioral traits can be used to distinguish individuals. Over time, as mathematicians and engineers developed better techniques for analyzing patterns, a new field of study emerged: biometrics.
Biometrics investigates ways of matching individuals to their physiological or behavioral features to produce accurate and automatic identification. There are at least nine different human features currently in use or being evaluated for biometrics: face, facial thermogram, fingerprint, hand geometry, hand vein, iris, retinal pattern, signature and voiceprint. Unlike Ali Baba-esque passwords or identification (ID) cards, biometric features do not rely on knowledge or tokens possessed by individuals. Instead, biometrics allows us to rely on natural traits to authenticate ourselves much more conveniently.
Far from being merely hypothetical, biometrics have already been extensively used around the world. Tens of millions of notebook computers include embedded biometric sensors to protect the owner’s sensitive or personal data from unauthorized access, while airports such as Singapore Changi Airport and Hong Kong International Airport have implemented fingerprint scanning technologies to facilitate personnel clearance.
“City planners around the world are designing cities that use biometric data for purposes such as personal safety and transport efficiency. The opportunities for technology breakthrough in the public safety field are immense,” said Dr. Paul Wang, chief technology officer of NEC Corporation’s Global Safety Division.
Figuring ID out
Biometric systems rely on four basic components: capture, reference databases, matching and action. In this model, the system’s sensor captures and stores biometric samples from an individual to reference later. This is known as the enrollment procedure. Next, each sample is compared to those enrolled in the reference database by a “matcher.” Lastly, a systematic decision-making process typically algorithmic but potentially involving human input as well—determines whether the match is true or false. This is usually followed by some action, such as sounding off an alarm or granting access into a facility.
Questions arise right from the preliminary stage of capture. Even with the most adaptive sensor technology, there is a chance of mismatch. This is because the matching process relies on the use of probabilistic algorithms and is reported on a scale of 0 to 1—with 0 being zero chance of match and 1 being identical match. Typical biometric technologies identify an underlying numerical threshold above which all matching outcomes are interpreted as true, below which they are interpreted as false. Hence, there is always a chance for mismatch, be it false acceptance (Alice is identified as Bob) or false rejection (Charlie is wrongly rejected).
Problems also exist at the user level. Firstly, human beings age and their faces grow old. Secondly, no two images of the same thumb are identical, by virtue of the continual changes in dirt and sweat patterns. Individuals may put on headgear or accessories that give rise to a significantly different look. Effective biometric technology must therefore be able to identify invariant “inner characteristics” of the biometric features that hedge against changes at the superficial level.
Furthermore, security becomes a critical issue with the widespread use of biometrics. We unknowingly leave traces of our biometric data in public spaces. From fingerprint marks on escalators to unconscious “photobombing” that gives strangers access to our faces, our biometric features are always susceptible to exploitation by others. Individuals seeking personal benefit may be able to capture and reproduce personal biometric features without consent.
Aware of this issue, organizations usually deploy security personnel to ensure that humans are going through security by legitimate means. But the threat still exists and it is the duty of law- and policy-makers to adopt robust strategies to effect safer biometric systems.
Hello, Big Data
Like all databases, the reference database is prone to access attempts by third-party hackers. This is especially so as we progress into an age of Big Data. The benefits of mining large datasets have encouraged many to advocate cross-platform and crossdisciplinary data sharing. Yet, as Mr. Julian Ashbourn, a respected technology consultant in biometrics, points out, this creates an even larger risk for biometrics database security.
“There will be much we can achieve via the intelligent exploitation of data. However, our approach should be a combination of absolute clarity of purpose coupled to an in depth knowledge of root level data and its inherent quality, matched with an equal knowledge of writing intelligent rules around the extraction of such information,” wrote Ashbourn on his website, for a discussion paper entitled Biometrics and Big Data.
Scientists are currently developing ways to overcome these challenges and make biometrics technology more secure. In the US, Professor Anil Jain of Michigan State University, a wellregarded biometrics expert, is looking into multimodal biometric systems—systems detecting multiple biometric features at once—for public and commercial contexts. Other biometrics companies are using sophisticated statistical tools to develop robust 3D face recognition algorithms to replace the widely used 2D ones.
Undoubtedly a hot research topic, biometric technologies will continue to evolve, improving existing platforms and possibly even changing our concept of identity. After all, in some ways each of us possesses our own biometric procedures to distinguish our friends and loved ones—people with whom we share special connections. Figuring out what makes each of us unique is as much a goal of science as it is of any individual.
This article was first published in the print version of Asian Scientist Magazine, Jul 2014.
———
Photo: Shutterstock.
To read more, subscribe to Asian Scientist Magazine in print and receive four issues of Asian Scientist Magazine delivered directly to your mailing address for 12 months, inclusive of taxes and postage.
