Combining palmprint & Finger-Knuckle-Print for user identification
Othaila Chergui, Hakim Bendjenna, Abdallah Meraoumia, Salim Chitroub · 2016
Biometrics is a field of analysis of biological characteristics of an individual. Its purpose is to recognize and automatically verify the identity of a person based on physiological or behavioral characteristics. Multi-modal biometrics which combining several biometric based systems, are increasingly studied. Indeed, they reduce some limitations of unimodal based biometric systems, such as the inability to acquire data of individuals or intentional fraud, while improving the recognition performance. These benefits of multi-modality with unimodal biometric based systems are obtained by fusing multiple biometric traits. In this paper, PaLMprint (PLM) and Finger-Knuckle-Print (FKP) modalities are combined to get the best of both worlds. They are recent and very important biometrics modalities due to there discriminatory power, robustness over time and there acceptability by users. To use, there characteristics must be extracted and enrolled for future comparison. The present proposed method is based on Histogram of Oriented Gradients (HOG) and Random Forest Transform (RFT) in order to improve the performance of the multi-modal biometric system based on PLM and FKP modalities. The results of the different classifiers are combined (fused) at the matching score level. In order to evaluate the proposed systems, a database of 165 persons is used. The obtained experimental results show the effectiveness and reliability of the proposed systems.