A biometric reader or scanning device is used to obtain the necessary verification data from a person. For example, in fingerprint biometrics applications, an optical sensor is employed to produce an image of the ridge structure at a fingertip, and this image serves as the basis for further access control activity.
The sensor unit forms the key interface between a user and a biometric access control system, making it important to minimize the rate of reading failure. The quality and usefulness of the data obtained through sensors often depends on camera characteristics because the majority of biometric data is composed of images, with the exception of audio-based systems, such as voice recognition, and chemically-based systems, such as odor identification.
In an access control system, the biometric data gathered by the sensor device must be evaluated to gauge whether it is suitable for processing. Typically, an algorithm designed for signal enhancement is applied to the data in order to improve its quality, but if the quality is insufficient for processing, the user may be asked to resubmit the data. After processing, a specific set of features is selected from the overall data set to represent the qualifying identity trait. In fingerprint scanning, the relative positions of small ridge points can be extracted for use as a biometric measure. The feature set used for assessment and extraction is known as the biometric template, and is stored within the system’s database.
After the feature set has been extracted from the gathered data, it is compared to the stored templates and matched with any identical points. The number of matching points between the input and the template provides a match score, which can fluctuate between readings depending on the quality of collected data. The matching device in a biometric system usually includes a decision-making apparatus that relies on the match score to either confirm a person’s identity or to determine the identity by correlating the score to a ranked list of possible identities stored in the database.
A biometric system database stores all the information needed for processing biometric readings. When establishing access control parameters, the feature template is input into the database, sometimes along with biographical information specific to the user in order to enhance security levels. The parameter setting and data gathering processes may be automated or under a technician’s supervision. For example, when establishing access control for an individual computer, a user may input his or her own data to prevent others from using the computer’s resources. However, when setting biometric controls in a restricted facility, a user’s credentials may need to be authorized by a manager or input under supervision.
A single biometric sample is often sufficient for extracting a user template, but in some cases, a biometric system can process multiple samples to form a mosaic representation. Some systems also store multiple templates to compensate for data variations that can arise from a single user. This method is often found in facial recognition biometrics, in which several templates are employed to account for alterations in facial poses relative to the sensor.