Biometrics is one type of authentication method which recognizes humans by their body characteristics, such as fingerprints, irises and eye retinas, voice patterns, facial patterns and hand measurements. There are two main categories for biometrics characteristics: 1) Physiological have connection to “shape of the body,” such as fingerprints, face recognition, hand and palm geometry, iris recognition, and so on. 2) Behavioral have connection to “the behavior of a person,” such as typing rhythm, gait, and voice. It is important to know the difference between verification and identification in order to understand the function of different kind of biometrics method. Verification is to check the user if he/she matches his/her smart card, user name, or ID number whereas identification is to check the user by comparing to all record in a file.
There are a variety of biomertics methods, and the followings are examples.
Face recognition: face recognition is the way that identifies and verifies an individual automatically in order to give permission to securely access to a system, such as a computer, a mobile device, and online e-commerce. Computer technologies contribute to recognize the person’s face by digitizing and matching the face in the face database. Eigenface, fisherface, the Hidden Markov model, and the neuronal motivated dynamic link matching are examples of popular recognition algorithms, and more recently three-dimensional face recognition which has succeeded to improve accuracies has been developed. One of the disadvantages may be that facial recognition can’t effectively work under certain circumstances, such as poor lightning, sunglasses, long hair, face expression and so on.
Fingerprint recognition: fingerprint recognition is the way that identifies and verifies an individual automatically by using fingerprint collation. It is essential to know some kind of print patterns and the structure and properties of human skin for successful fingerprint collation. The arch, loop, and whorl which are basic pattern of finger print ridges and ridge ending, bifurcation, and short ridge (or dot) which are the major Minutia features of fingerprint ridges are primarily used for fingerprint identification. Moreover, there are several types of fingerprint sensors, such as optical, capacitive, ultrasound, and thermal for capturing the digital image of fingerprint patterns.
Hand geometry: hand geometry is the way that verifies users by shape of their hands, but not identifies them. The reason for this is that people’s hands aren’t unique enough for identifying an individual compared with fingerprints. Consequently, hand geometry may be ideal for immigration and border control, since it is less likely to violate people’s privacy. Other benefits are easy to collect data of hand geometry and combine with other biometrics, like fingerprints. However, there is also a drawback that is to cause change of hands by injury, loss or gain weight, and arthritis over time. In the real world, Disney theme parks turn hand geometry to good account by finger geometry reader in order to admit grant ticket holders to different parts of the park.
Iris recognition: iris recognition is the way that verifies and identifies individuals by pattern recognition techniques of individual’s irides. Retina recognition (also called retina scanning) is similar to iris recognition in a way that catches a high quality picture of iris or retina by a digital camera which commonly requires illumination for obtaining image of iris or retina. Consequently, NIR (near infrared) light applies to both techniques. In addition, LED technology contributes to prevent from eye damage, but multiple illuminators may be harmful to eyes if they are designed and utilized without care. Advantages of iris recognition are less possibility of damage because of high protection with membrane (the cornea), easier prediction of shape of iris because of its simple structure, free from touch that often cause rejection from people and so on whereas disadvantages of iris recognition are high start up cost, less reliability compared with other photographic biometric technologies, less useful compared with fingerprints and DNA for forensic identification, and so on.
Speaker recognition: speaker recognition is the way that automatically recognizes the speaker by his/her voice in speech waves. Speech recognition is mistakenly used as synonym of speaker recognition by many people, but the two terms are different. In speech recognition words (“what is being said”) are recognized instead of “who is speaking”. Moreover, it is important to understand a distinction between the act of verification and identification. Verification or authentication applies if the purpose of speaker is to know certain identity and the voice is implemented for the purpose of a verification whereas identification applies if the purpose of a speaker is to know uncertain identity. There are two types of techniques to recognize a speaker: 1) text dependent (constrained mode) is the system that uses password or prompted phrase from users. 2) text independent( unconstrained mode) is the system without program of phrasing and more flexible way than text dependent.
Benefits of biometrics
1) Biometrics enables faster and easier authentication and allows user to recognize identity more smoothly than other types of authentication method, such as user ID, password, a smart card, or tokens. Moreover, it enormously contributes to information technologies and organizations for helping management of user authentication.
2) Unlike cards or password, biometrics reduces a possibility of stolen or lost identifier and provides more secure logical and physical access for users.
3) Initially, high cost of biometrics was a major concern, but in recent years, it has largely went down along with advancement of hardware and software technologies.
Drawbacks and issues of biometrics
1) Privacy is a major concern for users, since clear guidance for how individual’s identifier is securely stored and used hasn’t developed enough.
2) Log-lasting performance is a major advantage for biometrics, but it is also a huge issue for biometrics, since it is hard to cancel and reissue once individual’s identifier is collected.
3) Biometrics may cause a risk to owners which may ultimately cause higher cost than secured cost for their property. In fact, the Malaysian gang group cut off the owner’s finger in order to steal a Mercedes S-class in 2005, since the car was secured by fingerprints recognition system.