Studies On Automatic Fingerprint Identification Technique

Fingerprint verification is one of the most reliable personal identification methods. An automatic fingerprint identification system (APIS) is widely employed. It plays a vital role in providing forensic evidences as well as in our everyday lives, for instance, in criminal identification, access control, and ATM card verification. Thus, to execute deep researches on fingerprint recognition is not only theoretically but also practically meaningful. Based on latest extensive discourses and technology journals in this field, the dissertation is trying to make studies on common fingerprint recognition, from algorithms on fingerprint preprocessing, to fingerprint feature extracting and fingerprint matching.The main topics of this dissertation are listed as follows:1) Algorithms on fingerprint preprocessing is first studied in unit 3. First of all, a point direction image is acquired by using discrete direction method. Then a block direction image is created by a combination using of the rectangular histogram method and the dynamic block-cutting method. Next, fingerprint image can be handled with a filter, which is a direction selective system. Then the gray image is transformed into a binary image through a one-by-one-point section-threshold binary algorithm. After that, a postprocessing is implemented on the given binary image. In the part of thinning algorithm, An improved thinning algorithm derivedfrom the OPTA is introduced in, for the purpose of producing high quality skeleton of fingerprints, on which the more accurate fingerprint feature location and type are based. As a result,pinpoint fingerprint matching can be achieved. Preferences in the preprocessing are decided by repeating experiments.2) Algorithms on fingerprint feature extraction are studied in unit 4. Minutiae can be extracted with exactitude by using 8-neighborhood cross-calculation algorithm, a kind of algorithms featured especially in its ability of anti-jamming.3) In unit 4, A new feature extraction postprocessing algorithm basded on ridge tracing is introduced. It can totally remove pseudo-structures like spurs, bridges and holes from a thinned fingerprint image. Although, it consumes much time in the step of feature extraction, it saves us a lot in the step of fingerprint matching on the other hand, and moreover, this postprocessing step largely improves both the precision and the efficiency of the matching. After this step, spurious minutiae is not a trouble to us anymore. While applied to those low quality images, its evident effect can be easily beheld. So far, a solid foundation for further fingerprint matching has been established.4) In unit 5, An alignment-based algorithms minutiae matching for fingerprint verification is studied. This matching algorithm can reduce differences during translation, rotation, and zooming under the noisy or distorted condition. To do this, firstly, select optimum matching-origin-pair in the polar coordinates, then the number of matching pairs can be acquired.The validity of methods presented above are all well corroborated by substantive experiments.