Biometrics:Fingerprint TechnologyBiometrics: Fingerprint TechnologyCalvin ShuehProfessor StampCS2654/5/04Living in the information age, individuals have vast amounts of information that they wish to keep private. Much of this information is protected by the use of passwords. Although this approach is satisfactory with most individuals, some seek moresecure methods. One approach is using characteristics of individuals as the form of authentication, known as biometrics. Biometric security is based on something you knowor have, and are. Fingerprints are the most common form of biometrics and have several measurable distinctive characteristics. The biometrics industry is growing fast because ofnew technology and the need for a more secure authentication system.Although many individuals feel that passwords are enough to protect our information, there are several problems1. People often forget passwords or worse, they can be stolen and then used by other individuals. A person might have several different passwords used for different applications. Passwords are often poorly chosen because individuals incorporate personal information or use common dictionary words.An alternative to passwords is using human characteristics for the purposes of identification; this is known as biometrics2. Because individuals have these distinctive features, they can be used as a form of identification3. According to a study, as many as 80% of the public has allowed a biometric feature to be recorded4. Although there are several human characteristics that can be measured for authentication including the face, eye, and voice, the fingerprint is the most commonly used characteristic5. Everyone is born with a fingerprint. They cannot be forgotten at home or left in the car. Fingerprints are the oldest form of biometrics that has been used successfully6. In the 14th century, parents in China used the fingerprints and footprints of their children as a form of identification7. Since then, fingerprints have been studied and theircharacteristics have been catalogued. Each individual fingerprint is unique; everyone hasan immutable fingerprint8. A fingerprint consists of several lines that produce patterns, called ridges, which can be used to verify and authorize an individual9.The most common system used to classify the ridge patterns in fingerprints are known as the Galton features10. There are six classes of patterns. These are known as arch, tented arch, left loop, right loop, whorl, and twin loop. Each pattern has it’s own distinct design that distinguishes them apart. The several features that are classified are known as minutiae11. These are the irregularities in the otherwise smooth pattern of ridges in a fingerprint. The minutiae include characteristics called the crossover, core, bifurcation, ridge ending, island, delta, and pore12.Figure 1 13The crossover pattern is created when two different ridges cross each other. The point in which swirls or other patterns often center around is known as the core. Bifurcation is the point at which one ridge separates into two separate ridges. A ridge ending is the end point of a ridge. An island is small ridge in the space between two otherridges and does not touch any other ridges. The space in between ridges where severalridges surround is known as a delta. Occurring inside ridges at steady intervals are pores14. A device is used to capture an image of the pattern of an individual’s fingerprint. There are two main technologies used to capture the image of the fingerprint. The first technology involves optical technologies using a prism in which a source of light is refracted15. Using this light, the device is able to take an accurate fingerprint image.The second technology used is capacitive-based semiconductors. The fingerprint is obtained by having the subject place the finger on a sensor chip. The chip then detects capacitance changes between the ridges and valleys between the chip and skin and uses this to construct an image according to the variance of voltages16. An example of an optical technology is the Advanced Minutiae Based Algorithm by Suprema Solutions. The algorithm uses two processes: the Feature Extractor and Matcher17. When this image is captured, the extraction of the minutiae is known as Feature Extractor; this serves as the core of fingerprint technology18. The Feature Extractor will often need to enhance the image prior analysis19. Much of the features of the fingerprint caused by noise, poor contrast, and dirt can be removed by using a noise reduction algorithm20. This image is then processed using proprietary algorithms extract the minutiae and store or compare this information against a previously stored fingerprintin the database. Two of the most frequent features of the minutiae used in applications are the characteristics at ridge endings and points of bifurcation21 22.Figure 2 23During extraction of the minutiae several points of interest will be discarded due to distortion and false minutiae. A minutia might be discarded because it is too close to another minutiae24. This is because research shows that are very rarely adjacent to each other. There are several factors of a fingerprint that can be legitimate but be discarded bythe extractor. If a finger has a scar that makes it appear that there is a ridge crossing several other ridges, this can result in a minutia that will be discarded. There will be between 30 and 60 minutia remaining after the extraction and cleanup25. After extraction, there are several methods to store the information resulting from the remaining minutiae. One method is to use the core as the center point with coordinates (0,0) and determine the location of the remaining minutiae around the core26. Another algorithm is to use the boundaries on the bottom and left as the axes for determining coordinates. Besides the coordinate, the angle of the minutia is also calculated and stored. The angle is calculated by using the horizontal line from the core using the right of the core as the zero degree angle27.The second part of the algorithm is the Matcher. This is the part of the applicationthat will try and match the fingerprint that was obtained against fingerprints in a database.There is often a trade-off between speed and performance. Being able to quickly and correctly identify a fingerprint in the shortest amount of time possible with the high accuracy
View Full Document