We examine the robustness of algorithm performance over multiple datasets collected with different sensors. This study provide insight as to whether an algorithm performance derived from traditional controlled environment studies will robustly extrapolate to more challenging stand-off/real-world environments. We argue that a systematic methodology is critical in assuring the validity of algorithmic conclusions over the broader arena of applications. We present a structured evaluation protocol and demonstrate its utility by comparing the performance of the open-source algorithm over seven datasets, spanning six different sensors (three stationary, one handheld, and two stand-off types). We also provide results for the ranking of the seven datasets measured by four performance metrics. Finally, we compare our protocol-based ranking with a parallel ranking based on an independent survey results from a collection of biometrics experts, with high correlation between the two rankings being demonstrated.
Proceedings Title: The IEEE Sixth International Conference on Biometrics: Theory, Applications and Systems (BTAS 2013)
Conference Dates: September 29-October 2, 2013
Conference Location: washington, DC
Pub Type: Conferences
biometrics, iris recognition, robustness, algorithmic robustness, multiple datasets, data diversity, performance evaluation