Computer Vision: Biometrics

Engineering Undergraduate Programs


Computer Vision: Biometrics

The field of biometrics is concerned with the science of determining or confirming an individual’s identity based on measuring their unique physiological or behavioural characteristics. Nowadays, biometric systems are  increasingly being deployed in relation to many different types of scenario and user group including, for example, forensic investigations (e.g. criminal identification, parenthood determination and missing children), government (e.g. passports and other travel documents and border crossing) and widely varying commercial applications (e.g. mobile phones, internet banking and computer login). However, these unique characteristics are likely to change with the natural ageing process (passage of time) and, as a result, developing biometric applications for long-term use becomes a particularly challenging task. In our research, we study two established and widely used biometric modalities, iris and signature, and explore quantitatively the results of a detailed investigation into the physical effects of ageing, the relationships between physical ageing and interrelated physical factors which have a bearing on how the impact of ageing can most effectively be investigated and understood, and show how these factors can be manipulated in order to guide practical implementation towards achieving more reliable performance.

Although biometric systems utilise information which is assumed to be unique to an individual, thus allowing the possibility of identifying the ‘owner’ of the data, soft-biometrics (such as age, gender, emotion and so on) are explicitly not unique, but are nevertheless characteristic of an individual, and therefore can also be a contributory source of supplementary identification information. Although falling short of enabling an absolute identity decision, such measures at the very least allow elimination of some possible individuals from consideration in an identification scenario, narrowing the required search space and could thus be of considerable potential importance in forensic investigations. In our research, we study and explore predictive capabilities of iris biometrics with respect to age and gender, and emotion (such as happy, sad, stressed) prediciton capabalities with respect to signature and handwriting biometrics.

Faculty Members:
Meryem Erbilek

Related Publications:

  1. Y.B. Ayzeren, M. Erbilek and E. Çelebi, “Emotional State Prediction From Online Handwriting and Signature Biometrics”, IEEE Access, vol.7, pp. 164759-164774, 2019.
  2. M. Da Costa-Abreu, M. Fairhurst, M. Erbilek, “Age predictive biometrics: predicting age from iris characteristics”, Iris and Periocular Biometric Recognition, IET, Chap. 10, pp. 213-234, 2017.
  3. M. Fairhurst, M. Erbilek and C. Li, “A study of automatic prediction of emotion from handwriting samples to support forensic analysis”,IET Biometrics, vol.4, no.2, pp. 90-97, 2015.
  4. M. Fairhurst, M. Erbilek and M. C. D. C. Abreu, “Selective review and analysis of ageing effects in biometric system implementation”,IEEE Transactions on Human-Machine Systems,  vol. 45, no.3, pp. 294-303, 2015.
  5. M. Erbilek and M. Fairhurst, “A framework for managing ageing effects in signature biometrics”,IET Biometrics, vol.1, no.2, pp. 136-147, June 2012.
  6. M. Fairhurst and M. Erbilek, “Analysis of physical ageing effects in iris biometrics”,Future Trends in Biometric Processing,  IET Computer Vision, vol.5, no.6, pp. 358-366, November 2011.