Distribution alignment for cross-device palmprint recognition

作者：

Highlights：

• We propose a novel and simple loss term to narrow the crossdevice gap in the pairwise level. And we use a ‘progressive target’ to guide the estimated histogram.

• We established a new cross-device palmprint benchmark to improve the study of cross-device RGB palm recognition base on existing datasets.

• Extensive experiments on the newly collected dataset and several existing palm recognition datasets verified the effectiveness of the proposed method.

摘要

•We propose a novel and simple loss term to narrow the crossdevice gap in the pairwise level. And we use a ‘progressive target’ to guide the estimated histogram.•We established a new cross-device palmprint benchmark to improve the study of cross-device RGB palm recognition base on existing datasets.•Extensive experiments on the newly collected dataset and several existing palm recognition datasets verified the effectiveness of the proposed method.

论文关键词：Palmprint recognition,Deep learning,Loss function,Biometric recognition,Person Reidentification

论文评审过程：Received 17 October 2021, Revised 8 June 2022, Accepted 25 July 2022, Available online 30 July 2022, Version of Record 4 August 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108942