Finger Vein Template Protection Based on Alignment-Robust Feature Description and Index-of-Maximum Hashing
Simon Kirchgasser, Christof Kauba, Yen-Lung Lai, Zhe Jin, Andreas Uhl
Abstract
Privacy preserving storage and secure processing of biometric data is a key issue that has to be addressed in finger vein recognition systems as well. Various template protection approaches originally proposed for well established biometric modalities have been adopted to the domain of finger vein authentication. However, these adopted methods have the disadvantage that they are not designed for finger vein patterns in particular and are thus, sub-optimal in several ways. In this study we propose an alignment-robust template protection scheme that is based on an efficient binary representation of finger vein patterns on the one hand and is further using the advantages of Index-of-Maximum (IoM) hashing to fulfill mandatory privacy and security based characteristics on the other hand. The proposed method is compared to block re-mapping and warping regarding recognition performance and is analyzed with respect to security and privacy aspects. The analysis is further extended to robustness against misalignment of the finger vein data and a combination of block re-mapping/warping and the proposed method is investigated as well.