Efficient Sparse Representation for Finger-Knuckle-Print Biometrics
The paper introduces a novel sparse representation model for Finger-Knuckle-Print (FKP) biometrics, leveraging its uniqueness, accessibility, non-abrasiveness, and rich texture for personal identification. The proposed algorithm constructs an over-complete dictionary from training samples, utilizes Local Binary Pattern (LBP) operator for feature extraction and dimension reduction, and employs smooth l0 norm for model solving, leading to accelerated recognition process and improved efficiency. Experimental results on FKP Database demonstrate competitive performance compared to state-of-the-art methods, indicating significant practical potential.
文件大小：227.35KB
评论区