A fingerprint reader is a biometric tool used to verify a person's identity. These devices generally operate by sensing a user's fingerprint, forming a digital image, and comparing features of the live print to those of a stored copy. Fingerprint readers are used in law enforcement, building access, and computer security. Different technologies used in fingerprint readers offer different strengths and weaknesses for identity authentication.
Human fingerprints have a number of identifiable features that can be used for comparison. The ridges on a fingerprint typically form one of three distinct patterns. Arches are ridges that enter the finger from one side, rise in the middle, and exit the finger on the other side. Loops involve ridges entering the finger, looping around, and exiting on the same side. Whorls are sets of ridges that form concentric circles.
Individual ridges can also show identifiable features, called minutiae. Ridge endings are ridges that suddenly end without touching adjacent ridges. Bifurcation is when a ridge splits into two. A short ridge, or dot, is a ridge with a very short length. The number of possible fingerprint feature combinations is very large, and no two fingerprints have been shown to be a match.
A fingerprint reader uses a sensor to measure the fingerprint in question. Optical sensors record visible light in much the same way a digital camera does. This type of fingerprint reader typically has a specialized touch surface to flatten out a finger's curved shape. Optical sensors require both the touch surface and the user's finger to be clean in order to form a useful image. This type of fingerprint reader can also sometimes be fooled by an image of a fingerprint instead of a user's real print.
Ultrasonic sensors are another type of sensor used in fingerprint readers. They bounce high-frequency sound waves off the dermal layer of the finger to produce an image. The dermal skin layer shows the same fingerprint pattern as the outer skin surface, so ultrasonic sensors do not require clean touching or finger surfaces.
Once fingerprint data is gathered by a sensor, it is stored in a computer in an image file. An algorithm is generally used with a fingerprint reader to automatically recognize ridge patterns and minutiae. For this to be possible, the algorithm must first align the candidate and stored image in the same orientation. It can then measure relationships between features in the fingerprint, such as number of ridges between selected minutiae. In addition to algorithmic comparison techniques, an on-site staff may be called upon to visually inspect ambiguous fingerprint images for another level of verification.