상세정보
학술지
Multimodal Biometric Method that Combines Veins, Prints, and Shape of a Finger
- 발행일
- 201101
- 출처
- Optical Engineering, v.50 no.1, pp.1-13
- ISSN
- 0091-3286
- 출판사
- SPIE
- 초록
- Multimodal biometrics provides high recognition accuracy and population coverage by using various biometric features. A single finger contains finger veins, fingerprints, and finger geometry features; by using multimodal biometrics, information on these multiple features can be simultaneously obtained in a short time and their fusion can outperform the use of a single feature. This paper proposes a new finger recognition method based on the score-level fusion of finger veins, fingerprints, and finger geometry features. This research is novel in the following four ways. First, the performances of the finger-vein and fingerprint recognition are improved by using a method based on a local derivative pattern. Second, the accuracy of the finger geometry recognition is greatly increased by combining a Fourier descriptor with principal component analysis. Third, a fuzzy score normalization method is introduced; its performance is better than the conventional Z-score normalization method. Fourth, finger-vein, fingerprint, and finger geometry recognitions are combined by using three support vector machines and a weighted SUM rule. Experimental results showed that the equal error rate of the proposed method was 0.254%, which was lower than those of the other methods. © 2011 Society of Photo-Optical Instrumentation Engineers (SPIE).
- KSP 제안 키워드
- Biometric features, Fingerprint recognition, Fourier Descriptor, Multimodal Biometrics, Normalization method, Optical instrumentation, Principal Component analysis, Recognition Accuracy, Recognition method, Short time, Sum rule