This paper focuses on the recognition performance improvement of the template matching. In template matching, the probe image is accepted if its overlapped vein points with the enrolled image are more than the predefined threshold. However, the acceptance may be false as the structures of the probe vein pattern and the enrolled vein pattern may be different even with a large number of overlapped vein points. Therefore, the vein structure may have the capacity to enhance the template matching.
In this paper, the vein structure near the bifurcation point of vein pattern, named the tri-branch vein structure, is explored and employed to improve the performance of template matching by the proposed user-specific threshold based filter framework. The tri-branch vein structures are extracted from the whole vein pattern with the help of vein network. In the first level of the framework, the structures are used to filter some imposter images by user-specific threshold (i.e., the minimum genuine score of each user). In the second level, the whole vein pattern is employed to give the recognition result in the reduced matching space.
