Seven Pillars of Biometrics (Part 2)

TalkingBiometrics@ITREALMS:

In this edition, we would continue discussion on the characteristics that determine the suitability of a human trait as a biometric attribute, otherwise called the seven pillars of biometrics.

4: Circumvention or Security

A measure of how easy it can be to fool a given attribute is an indication of its suitability (or otherwise) for biometric purposes. Circumvention is a quality that evaluates the ease with which a biometric trait can be fooled or bypassed by an attacker and to what extent it can withstand being spoofed or defeated. The measure or degree of difficulty required to bypass a biometric trait is an indication of its security and usability as a biometric authentication technique, and also a pointer to how vulnerable it might become to spoof attacks and identity fraud.

It is common occurrence for two separate individuals to have the similar height or exactly same weight, but it is rare to obtain a perfectly similar facial pattern or fingerprint from two persons. This is why height and weight cannot be used as classical biometric attributes, but facial pattern and fingerprints can be used because it is difficult to obtain such similar values from different persons, hence they are more secure. It is also more difficult (but not totally impossible) to fabricate fake eyeballs or retina patterns than to obtain fake signatures. 

In this sense we say that even though both offer relatively high levels of security as biometric attributes, retina patterns are more difficult to circumvent or forge than signatures, thereby making retina patterns a more secure set of biometric attributes than signatures, and a more preferred option for use at maximum security facilities. Vein patterns have also been known to exhibit a more secure biometric characteristics than voice prints.

5: Performance

Biometric performance is a measure of the speed and accuracy of recognizing the trait while in normal use even in the midst of environment effects. It is an indication of how well the biometric trait is able to compensate for inherent losses during the processing of the samples. In general biometric performance is closely tied to its efficiency to successfully complete the identification or authentication processes in a timely manner while minimizing errors. 

As an illustration, it is faster and more probable to process 3D facial patterns from babies and infants than to recognize their fingerprints. This is because fingerprint formation naturally takes time to become fully-usable as a biometric trait, resulting in the poor quality of fingerprint processes often experienced with toddlerbiometrics. In this sense, we say that facialprints have better infant-related performance than fingerprints.

… To be continued next week.

ITREALMS ... everything news digitally!

About The Columnist

Kenneth Okereafor is an accomplished ICT professional and Cybersecurity expert with over 20 years computing experience in Enterprise ICT innovation management in both private and public sectors. He holds dual PhD degrees in Computer Science (Cybersecurity & Biometrics) and Management (ICT Administration and Governance) with a first class Electronics & Telecommunications Engineering background. He is a United Nations trained Network Security Specialist and is proficient in Network Optimization, Data Protection, Biometric Systems and Digital Forensics. In addition to many published works and conference papers, Kenneth has participated at many international technology events, and has contributed to several national ICT development programmes including the design and development of the National eHealth Strategic Framework 2015 - 2020. He has strong research interests in Digital Innovations, Cloud Security, Cybersecurity Education, Biometric Liveness Detection and Green Computing; and may be reached at nitelken@yahoo.com