From time to time issues arise within the research community which spark widespread debate and, sometimes, disagreement.
But such discussion is to be encouraged, not least because it can lead to greater light being shed on the relevant issue and, ultimately, to greater insight into and understanding of thorny problems.
In the biometrics research community there is just such an issue which has been generating considerable debate in recent months: the question of “iris ageing” - specifically, the effect of the passage of time on the recognition of individuals based on the analysis of their iris characteristics. In response to this debate, and in the spirit of encouraging productive and enlightening discussion, the international research journal IET Biometrics is publishing, in the same issue (December 2015), two key papers in this area from authors who have rather different views on the topic. One of these contributions (written by Ortiz and Bowyer of the University of Notre Dame) is “A critical examination of the IREX VI results”, which reviews a report published originally by the US National Institute of Standards and Technology (NIST). The original NIST report looked at 3.5 million iris images from over 600,000 persons collected in an operational system over a period of about 6 years and found no evidence of an iris aging trend. The Ortiz/Bowyer paper argues that improvements in the methodology used in the NIST report would lead to an uncovering of an ageing trend leading to a decrease in recognition rate over time.
The second paper (written by Grother and Matey, two authors of the original NIST report) is “IREX VI: mixed-effects longitudinal models for iris ageing”, which argues that the large ‘ageing’ effects claimed by UND in their series of papers, which led to headlines in the popular press, are the result of lack of control of ambient conditions during their (Notre Dame’s) data collections and do not represent changes in the underlying iris pattern.
This is an area of such importance and current debate that it is considered timely to make these two papers, together with a brief scene-setting Guest Editorial by Professor James Wayman, more generally available to the wider scientific community, as a means of furthering our understanding of this issue and encouraging better informed debate.
Professor Michael Fairhurst, IET Biometrics Editor-in-Chief