Finger-knuckle-print Recognition Based on Image Sets and Convex Optimization

In this paper, A finger-knuckle-print (FKP) recognition algorithm based on image set and convex hull optimization model is presented. The proposed method start from the viewpoint of inputting biometric image sets, and seeking for the suitable convex hull optimization model. Local phase quantization is utilized for feature extraction, and the extracted feature is adopted for the convex hull model construction and optimization, in order to finish finger-knuckle-print recognition. Simulation experiments show that, the proposed algorithm can achieve good performance on the public FKP database.