Study On Microcomputer Fingerprint Recognition Technique

Biometric recognition technique, which is represented by fingerprint recognition technique, will be used more and more widely in many fields. At the same time, fingerprint recognition technique will play a dominant role in identity recognition. Eventually it will become an indispensable part of our daily life. Therefore, studies on the theory and application of fingerprint recognition have significant meanings both in theory and application.Based on plenty of papers, technology reports and dissertations, this dissertation makes some research on common fingerprint recognition, which including algorithms on fingerprint preprocessing, fingerprint feature extraction and fingerprint matching.The main points of this dissertation are as follows: 1. Algorithms on fingerprint preprocessing are first studied in unit 3. Point direction image is acquired by using discrete direction method. Then block direction image is acquired. Fingerprint image can be filtered by a filter, which is changed by direction. Then gray image is changed to binary image with a local smoothing self-adapted binary algorithm. In the part of thinning algorithm, we introduce a two-pass sequential iterative thinning algorithm, which produces high quality skeleton of fingerprints. Fingerprint feature location and type can be more accurately acquired with this thinning algorithm. Thus, precision of fingerprint matching can be improved. Some parameters in Preprocessing are selected by repeating experiments.2. Algorithms on fingerprint feature extraction are studied in unit 4. Minutiae can be extracted using our template-matching algorithm. And minutiae extraction algorithm is robust. 3. Fingerprint feature validating Algorithm is studied in unit 4. A new feature-validating Algorithm is introduced. Although, validating Algorithm need more time in feature extraction step, but this validating step save a lot of time in fingerprint matching step. At one time, this validating step improves the matching precision. After this step, spurious minutiae can be eliminated effectively. The preserved numbers of true minutiae is 1/4 to 1/3 of pre-validating. As for some low quality image, effect is more obvious. And the rate will be 1/10. A nice foundation is established for fingerprint matching.4. In unit 5, Algorithms on fingerprint matching is studied. We have developed an improved feature point-matching algorithm. Then matching of point is transformed into matching of vectors, which are composed of two feature points. This can eliminate the influence from translation along x and y. This matching algorithm can handle translation, rotation, and scaling differences under noisy or distorted condition. Thus matching pairs can be acquired.