Multimodal finger-feature fusion and recognition based on tolerance granular space

Abstract

A finger has three modalities, fingerprint (FP), finger-vein (FV) and finger-knuckle-print (FKP). Taking these finger traits as a whole for recognition is a very natural maneuver. Granular computing can effectively solve a fusion problem using knowledge from multiple levels of information granularity. Viewing finger-based recognition as a multi-granularity problem, a multimodal finger feature tolerance granular space model (MFTGSM) is proposed to implement feature fusion of FP, FV and FKP. Granular space is constructed in bottom-up manner, and a sliding window scheme is adopted for processing a multilevel granular partition with suitable overlapping. For saving computational cost, based on MFTGSM, the recognition process is implemented using a coarse-to-fine granular matching strategy. Experiments are performed on a self-built database with the three modalities. And the results demonstrate that the proposed method achieve good results in identification performance with higher reliability and accuracy.