Local descriptors in application to the aging problem in face recognition

Abstract

Local descriptors are widely used in face recognition due to their robustness to changes in expression or occlusion in facial images. In this paper, a comparison of local descriptors commonly used in face recognition methods is presented in the context of age changes of individuals. We quantify abilities of local descriptors used in face recognition in the context of age discrimination. The performance of the descriptors is evaluated by experimenting with the FG-NET database. We present the results for different age groups and for various age differences of individuals present in the training and testing images. The values of recognition accuracy are reported in combination with various similarity measures used for classification purposes. Moreover, the performance of the descriptors combined with Gabor wavelet images is tested. © 2013 Elsevier Ltd. All rights reserved.