Cancelable fingerprint template design with randomized non-negative least squares

作者：

Highlights：

• The non-invertible transformation of the PLS descriptor is formulated as a problem of solving randomized Non-Negative Least Square (R-NNLS) optimization problem, which is unique in a way to get rid of the performance deterioration issue when applying the non-invertible transformation directly to the minutiae descriptor for the fingerprint template protection.

• The revocability and the unlinkability traits of the proposed method are enabled by introducing a permutation mechanism to the R-NNLS.

• The proposed method is rigorously evaluated with five benchmark datasets, and four security and privacy analyses are also carried out.

摘要

•A novel alignment-free minutiae descriptor that based on Partial Local Structure (PLS) is proposed, which is beneficial to generate discriminative and compact binary cancelable fingerprint template.•The non-invertible transformation of the PLS descriptor is formulated as a problem of solving randomized Non-Negative Least Square (R-NNLS) optimization problem, which is unique in a way to get rid of the performance deterioration issue when applying the non-invertible transformation directly to the minutiae descriptor for the fingerprint template protection.•The revocability and the unlinkability traits of the proposed method are enabled by introducing a permutation mechanism to the R-NNLS.•The proposed method is rigorously evaluated with five benchmark datasets, and four security and privacy analyses are also carried out.

论文关键词：Cancelable biometrics,Partial local structure descriptor,Randomized non-negative least square,Binary representation,Fingerprint template protection

论文评审过程：Received 4 July 2018, Revised 4 December 2018, Accepted 25 January 2019, Available online 30 January 2019, Version of Record 5 March 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.01.039