How Fingerprint Recognition Works

Biometrics is the identification of a person based on his/her intrinsic characteristics, such as voice, writing, fingerprint, iris, face, hand geometry, etc. It is a system that recognizes a person based on "who" the person is and does not rely on "what a person is carrying" or "what a person knows".

One of the most widely known and used biometric identification scheme is fingerprint. Fingerprint analysis and identification is a 100-year old science, with thousands of studies made and articles written, and thus confidence rates of more than 99%. It's the cheapest, fastest, most convenient and most reliable way to identify someone. That's why fingerprint alone has almost 60% of the biometric world market, according to the Biometrics Market and Industry Report 2007-2012 from the International Biometric Group. Cars, cell phones, PDAs, personal computers and dozens of products and devices are using fingerprints more and more. Fingerprint recognition can be used from criminal investigations to time attendance systems, in a movie rental or in a police identification institute.

A fingerprint contain a number of unique physical characteristics called minutiae, which includes certain visible aspects of fingerprints such as ridges, ridge endings and bifurcations (forks in ridges). Minutiae are generally found in the core points of fingerprints, located near the center of the fingertips. These characteristics are used to distinguish two fingerprints, or to state that they are the same. Even identical twins have different fingerprints.

There are a number of different ways to get an image of a fingerprint, and the most common methods today for electronic fingerprint readers are optical scanning and capacitance scanning. When a fingerprint is applied to - or passed over - the sensor window of the fingerprint reader, the fingerprint is scanned and a gray-scale image is captured. A special computer software then identifies the key minutiae points from the image. These points are then converted into a unique digital representation, called "template", comparable to a very big password.

When a fingerprint enrollment is done, only the fingerprint template is stored, not the actual image of the fingerprint. The fingerprint template is not only smaller than the finger image, but also faster to process when comparing two fingerprints.

Fingerprint recognition technology is divided into two distinct processes: verification and identification.

In the verification process the user states who he/she is and a fingerprint is taken and compared to the user's previously registered fingerprint. If the fingerprints match, the user is "verified" as who he/she says he/she is. Since the newly acquired fingerprint is compared to only one stored fingerprint, this is called a one-to-one matching process (1:1). As in the enrollment process, when a fingerprint verification is done, only the fingerprint template is used in the comparison, not the actual image of the fingerprint.

In the identification process the user doesn't need to state who he/she is. A fingerprint is taken and compared to each fingerprint in the database of registered users. When a match occurs, the user is "identified" as the existing user the system found. Since the newly acquired fingerprint is compared to many stored fingerprints, this is called a one-to-many matching process (1:N). As in the verification process, when a fingerprint identification is done, only the fingerprint template is used in the comparison, not the actual image of the fingerprint.

Although it may appear to be a simple process, fingerprint recognition involves the state-of-the-art technology in hardware, for the readers, and software, to process the fingerprint images and templates. But due to the effort of the hardware and software makers, all this complexity is hidden and fingerprint technology usage is nowadays as simple as touching the reader.