An important step in finger-vein recognition is a proper alignment of the finger vein patterns to be compared. This alignment method is used to handle finger pose variation and to enhance the stability of the finger vein authentication. We proposed the iterative closest point (ICP) method on finger contour for the alignment. After the aligning process, we applied maximum curvature as a vein feature for the comparison without further optimization of the parameters, and we gained a better genuine comparison score. Even though this method is robust to finger pose variations, we have to further verify whether it will impact recognition performance. The experimental results show that the accuracy of the proposed method in verification case is enhanced and increases slightly than the state-of-the-art registration method which is called center line registration. Using ICP for registration resulted in a reduction of the EER to 0.3% (from 0.7% for the centerline registration) and a reduction of the FNMR@FMR0.01 to 0.7% (from 2.0% for the centerline registration).
|Title of host publication||2019 Symposium on Information Theory and Signal Processing in the Benelux|
|Editors||Gilles Callebaut, Kevin Verniers, Bert Cox|
|Place of Publication||Leuven|
|Publisher||Katholieke Universiteit Leuven|
|Number of pages||5|
|Publication status||Published - 28 May 2019|
|Event||40th WIC Symposium on Information Theory in the Benelux 2019 - Ghent Technology Campus, Leuven, Belgium|
Duration: 28 May 2019 → 29 May 2019
Conference number: 40
|Conference||40th WIC Symposium on Information Theory in the Benelux 2019|
|Period||28/05/19 → 29/05/19|
Normakristagaluh, P., Spreeuwers, L., & Veldhuis, R. N. J. (2019). Finger-vein Pattern Recognition Based on ICP on Contours. In G. Callebaut, K. Verniers, & B. Cox (Eds.), 2019 Symposium on Information Theory and Signal Processing in the Benelux (pp. 97-101). Leuven: Katholieke Universiteit Leuven.