Manual forms of fingerprint identification procedures for people have been available for about 100 years and are mainly used in the area of criminal prosecution. Today, in its IT-supported, automated form, the biometric digital fingerprint identification procedure has a high recognition performance. Special optical, capacitive (semiconductor), thermal or direct-optical sensor technology is used to record the fingerprint in automatic fingerprint recognition. Ultrasonic sensors that measure the acoustic resistance of the skin, which differs between ridges and furrows, are still in the testing phase.

Regardless of how the fingerprint is recorded, the procedure always produces a grayscale image of the finger i.e. the fingerprint. This image is processed further so correct matching results can be achieved with the enhanced image. Stages in the image processing include reducing the image noise, enhancing the image and detecting the characteristics.

Extracting the characteristic features from the image can be performed using different methods. Either the entire image (global pattern matching), relevant parts of it, or the intricate details according to type, position and direction can be recorded. By comparing the measured characteristic features with the saved target values, the technology shows whether the prints come from the same finger and thus the exact same person.

A general overview of fingerprint recognition is available here (PDF).

An overview of the history of fingerprinting is available here (PDF).