LAUSR

In practical applications, the biometric authentication system requires a lower error rate, higher speed, and smaller template size. To satisfy such requirements, we propose a robust and efficient finger vein recognition method, consisting of three stages: preprocessing, feature extraction, and comparison. During the preprocessing, geometric and photometric variations are suppressed. The feature extraction is based on a high-level descriptor derived from the binarized difference of Gabor jet (BDGJ) extracted from predefined grid points in the finger region. In the descriptor, the feature vectors at the vertices of the grid are weighted according to their position relative to the finger region. This descriptor provides a high discrimination ability with robustness. The comparison is conducted in a coarse-to-fine approach based on the Hamming distance, with remaining horizontal shifts and finger rolling compensated. The experimental results show that the proposed method has state-of-the-art performance and efficiency for practical applications. In particular, the method’s equal error rate (EER) is 0.00% on some famous public databases, and its computational cost is so low that the method can be run on a mobile processor with high speed.