Multimodal biometric recognition system using SURF algorithm

Abstract

Personal identification systems based on biometrics are employed in many applications to enrich the security of person identification systems. A number of biometric traits are used in various applications. Among the different biometric traits, the palm vein images are used for person recognition because of its uniqueness. To improve the recognition rate of the person authentication system, we proposed the multimodal biometric recognition system, in which the palm print and palm vein images are fused together. Initially the captured palmprint and palm vein images are enhanced by Adaptive Histogram Equalization. Then the features are extracted from the palm print and palm vein images using Speeded Up Robust Feature algorithm. These extracted features are stored as feature vectors in the database. These feature vectors are combined using simple sum rule. Finally, the Sum of Absolute Difference distance measure is used to find the similarity between the test and database feature vectors.