A novel classification-selection approach for the self updating of template-based face recognition systems

作者：

Highlights：

• A classification/selection approach with a maximum number of p templates per user was proposed in order to keep limited the storage and computational requirements;

• Three facial data sets were used as test;

• We evaluated the performance of the proposed and of the other state-of-the-art methods using hand-crafted (BSIF) and auto-encoded (FaceNet, ResNet50, SeNet50) features;

• The proposed method showed very good performance. The face recognition performance is superior to that of other state-of-the-art approaches.

摘要

•A classification/selection approach with a maximum number of p templates per user was proposed in order to keep limited the storage and computational requirements;•Three facial data sets were used as test;•We evaluated the performance of the proposed and of the other state-of-the-art methods using hand-crafted (BSIF) and auto-encoded (FaceNet, ResNet50, SeNet50) features;•The proposed method showed very good performance. The face recognition performance is superior to that of other state-of-the-art approaches.

论文关键词：Self-update,Face recognition,Adaptive systems

论文评审过程：Received 6 May 2019, Revised 15 October 2019, Accepted 19 November 2019, Available online 27 November 2019, Version of Record 28 November 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.107121