Using a Multi-Instance Enrollment Representation to Improve 3D Face Recognition

One of most challenging problems in 3D face recognition is matching images containing different expressions in the probe and gallery sets. Face images containing the same expression can be accurately identified; however, realistic biometric scenarios are not guaranteed to have the same expression in both probe and gallery. In this paper we examine a multi-instance enrollment representation as a means to improve the performance of a 3D face recognition system. Experiments are conducted on the ND-2006 data corpus which is the largest set of 3D face scans available to the research community. In addition, we show that using a gallery comprised of multiple expressions offers consistently higher performance than using any single expression.