A multi-objective optimization approach for palm vein feature selection based on the discrete PSO

palm vein biometrics has received huge interest from the biometric research community because of its uncountable advantages. However, the feature extracted from the palm vein images may contain noisy, irrelevant, and redundant features which imply the decrease in performance of the palm vein biometrics system and also lead to the dimension problem. To overcome these problems, we present in this paper a multi-objective optimization approach for palm vein feature selection based on the discrete Particle Swarm Optimization (PSO). The proposed multi-objective function takes into consideration both the accuracy of the biometrics system and the size of the selected feature vector aiming at maximizing the accuracy of the system and minimizing the size of the feature vector. The proposed approach has been tested on the CASIA palm vein database and compared to the full feature-based verification system and the discrete PSO-based feature selection with the traditional objective function. The experimental results show that the proposed multi-objective function outperforms the state-of-the-art.