This paper addresses the problem of developing facial image quality metrics that are predictive of the performance of existing biometric matching algorithms and incorporating the quality estimates into the recognition decision process to improve overall performance.
The first task the authors consider is the separation of probe/gallery qualities since the match score depends on both. Given a set of training images of the same individual, the authors found the match scores between all possible probe/gallery image pairs. Then, they defined symmetric normalized match score for any pair, modeled it as the average of the qualities of probe/gallery corrupted by additive noise, and estimate the quality values such that the noise is minimized. To utilize quality in the decision process, the authors employed a Bayesian network to model the relationships among qualities, predefined quality related image features and recognition. The recognition decision is made by probabilistic inference via this model. The authors illustrate with various face verification experiments that incorporating quality into the decision process can improve the performance significantly. (Publisher abstract provided)