Sparse Representation for Accurate Person Recognition Using Hand Vein Biometrics

The sparse representation theories are emerging as a more elegant and powerful technique to represent and analyze the biometric samples. In this paper, we study the feasibility of sparse representation on hand vein biometric data. Since hand vein data consists of rich set of textures, we first represent this texture information using Gabor transform. We then employ the sparse representation classifier to accurately classify this texture information to accurately recognize the individual using hand vein biometrics. Extensive experiments are carried out on public available hand vein data set of 100 users. Finally, the efficacy of the proposed scheme is also validated on the low quality (noisy) hand vein samples.