Biometric Algorithm Research Articles

Nonparametric measures of dependence for biometric data studies

This paper explores the possibility of using nonparametric dependence characteristics to evaluate biometric systems or algorithms. The extensions of classical rank correlation coefficients to the case when only a given number of top matches is used, are investigated. Difficulties with these coefficients in capturing the total correlation are noted. A version of a scan statistic, which measures co-occurrence of rankings for two arbitrary algorithms, is studied. The exact covariance structure of this statistic is found for a pair of independent algorithms. In the general case, its asymptotic normality is derived by using classical results on linear rank statistics. The concept of copula is shown to be useful for the study of nonparametric dependence characteristics. The results are applied to an example from the FERET (Face Recognition Technology) program. It is demonstrated that the random scores of considered recognition methods can be modeled by a two-parameter family of copulas exhibiting strong tail dependence.

Ultrasonographic prediction of term birth weight: How accurate is it?

Objective: The purpose of this study was to investigate the accuracy of 25 ultrasonic algorithms for the estimation of term fetal weight and to compare these results to an equation that is based on maternal and pregnancy-specific characteristics alone. Study design: Ultrasonography was performed in 82 nondiabetic gravid women at 35 to 41 weeks of gestation. Fetal biparietal diameter, abdominal circumference, head circumference, and femur length were measured. Twenty-five ultrasonic fetal biometric algorithms and an equation that is based only on maternal characteristics were used to predict birth weight. Results: The correlation between predicted and actual birth weight ranged from 0.44 to 0.79 for the ultrasonic algorithms and was 0.60 for the maternal characteristics equation. Ultrasonic algorithms had mean absolute prediction errors that ranged from ±263 to 646 g (±7.5%−18.8%). Accuracy for the maternal characteristics equation was not statistically different from the best performing ultrasonic algorithms (±353 g, ±10.4%). Conclusion: Term birth weight estimates that use ultrasonography are generally no more accurate than predictions that are based solely on quantitative assessment of maternal and pregnancy-specific characteristics. (Am J Obstet Gynecol 2003;188:566-74.)