Biometrics is an emerging technology for automatic human identification and verification using unique biological traits (Woodward, Orlans, & Higgins, 2002). These traits include face, fingerprints, iris, voice, hand geometry, handwriting, retina, and veins. For example, fingerprint recognition analyzes ridge ends, bifurcation, or dots of finger tips; voice recognition analyzes speech signal characteristics; iris recognition analyzes the pits, striations, filaments, rings, dark spots, and freckles of eyes; and face recognition analyzes facial parameters (Du et al., 2004). It is based on “something you are” rather than “something you have” (Du, 2005). Compared to the traditional identification and verification ways, such as user name/password, and paper IDs, biometrics is more convenient to use, reduces fraud, and is more secure (Reid, 2004).
Biometric system usually includes two subsystems: (1) the biometric enrollment system (Figure 1a) and (2) biometric matching system (Figure 1b).
The biometric enrollment system includes the Sensor Module, the Data Acquisition Module, the Data Preprocessing Module, the Pattern Analysis Module, the Pattern Extraction Module, and the Biometric Database Module. And the Data Acquisition Module interprets the biometric data into digital signals (images). The Data Preprocessing Module processes these signals to reduce the noise. The Pattern Analysis Module finds the most distinctive patterns of the biometric traits. The Pattern Extraction Module picks these distinctive patterns and generates identifiable templates. These templates will be then saved in the biometric database.
Compared to the biometric enrollment system, the biometric matching system adds the Pattern Matching Module and the Decision Module. In the biometric matching system, the newly sensed biometric data will be first processed similarly as the enrollment data, and the system will generate the pattern templates from the data. The Pattern Matching Module compares the newly generated templates with those in the biometric database and calculates match scores or quality scores for final decision. If the matching score is higher than the predetermined threshold, the system identifies/verifies it.
The false acceptance rate (FAR) and the false rejection rate (FRR) are used to measure if the biometric system is reliable (Ratha, Connell, & Bolle, 2001). A biometric system that generates high scores of either FAR or FRR is not reliable and cannot be used.
The FAR measures the percentage of incorrect identification:. (1)
The FRR measures the percentage of incorrect rejection:. (2)Top
Currently, the common used biometric systems include fingerprint, iris, face, and voice. The following paragraphs briefly describe each biometrics technology.
Key Terms in this Chapter
False Rejection Rate (FRR): FRR is the percentage of incorrect rejection. It is defined as the total number of rejection by the system divided by the number of false rejection.
Liveness Test: A test performed to test if the biometric traits are from a living person rather than an artificial or lifeless person.
Multimodal Biometrics: Automatic recognition of a person’s identity using more than one physical, biological, or behavioral characteristic.
Face Recognition: Automatic recognition of a person’s identity by mathematical analysis of person’s face features.
Biometrics: An emerging field of technology that uses unique physical, biological, or behavioral traits for automatic human identification and verification.
Voice Recognition: Automatic recognition of a person’s identity by mathematical analysis of person’s voice features.
Fingerprint Recognition: Automatic recognition of a person’s identity by mathematical analysis of the patterns of fingerprints.
False Acceptance Rate (FAR): FAR is the percentage of incorrect identification. It is defined as the total number of acceptance by the system divided by the number of false acceptance.