ABSTRACT
Biometrics technology has been attracting extensive attention from researchers and engineers of personal authentication due to the ever-growing demands on access control, public security, forensics, and e-banking. With the fast development of biometric data acquisition sensors and data-processing algorithms, diverse biometric systems have now been widely used in various applications. Among these biometric technologies, the handbased biometrics, including fingerprint, two-dimensional and three-dimensional palmprints, hand geometry or hand shape, hand vein, finger-knuckle-print, etc., are the most popular and have the largest shares in the biometrics market. This is due to the advantages of these traits, such as low cost, low-resolution imaging, and stable features. However, there are still many challenging problems in improving the accuracy, robustness, efficiency, and user-friendliness of hand-based biometric systems, and new problems are also emerging with new applications, for example, personal authentication on mobile devices and the Internet. Many types of unimodal biometric systems have been developed. However, these systems are only capable of providing
5.5.3 Unimodal System Identification Test Results 135 5.5.3.1 Open Set System Identification Palmprint 135 5.5.3.2 Closed Set System Identification Palmprint 136 5.5.3.3 Open Set System Identification Fingerprint 137 5.5.3.4 Closed Set System Identification Fingerprint 138 5.5.3.5 Open Set System Identification FKP 138 5.5.3.6 Closed Set System Identification FKP 139
5.5.4 Multiple-Instances System Identification Test Results 139 5.5.4.1 Fusion at the Matching Scores Level 139 5.5.4.2 Fusion at the Decision Level 144
5.5.5 Multimodal System Identification Test Result 144 5.5.5.1 Fusion at the Matching Scores Level 144 5.5.5.2 Fusion at the Decision Level 147
5.6 Comparing Results 147 5.7 Conclusion 148 References 148
low to middle range security features. Thus, for higher security features, the combination of two or more unimodal biometrics is required. In this chapter, we propose a multimodal biometric system for person identification using palmprint, fingerknuckle-print, and fingerprint modalities. This work describes the development of a multibiometric personal identification system based on minimum average correlation energy filter method (for matching). Therefore, a fusion process is proposed for fusing these traits. Comprehensive reviews on unimodal and multimodal biometric systems are given. We report on our latest results and findings in hand-based biometrics authentication, and we propose new ideas and directions for future development.
