Study Of Fingerprint Matching Algorithm Based On Local Structure And Qualityfactor

Fingerprint identification, which is an integrated technique using fingerprint for automatic recognition, is a research focus in the area of biometric recognition. Application of fingerprint identification has a long history. A lot of researchers are working hard on fingerprint identification technology and proposed a number of effective algorithms with fruitful results. However, there are still some problems unresolved. The expanding applications of fingerprint identification and the diversity of fingerprint acquisition instruments set higher requirements to the reliability and robustness of fingerprint matching algorithms. Otherwise, the quality of fingerprint image is also a critical factor to affect the accuracy of fingerprint matching. Therefore, this paper focuses on the identification of fingerprints with small size and little number of minutiae points. In order to improve the accuracy of the identification algorithms, we also take the quality of local area of fingerprint which around each minutiae into account in the matching process.For these fingerprints which have small acquisition area and with little number of minutia, normal matching algorithms based on point modal or ridge similarity have limitations. This work proposes a novel matching algorithm which is based on local structure of minutiae to deal with those cases. It uses effectively the information of limited minutia and its adjacent ridges. Firstly, the similar degree of ridges is used to seek the base points which are lied on those ridges with high similarity, then adjust the posture of the matching fingerprint based on those points. Sencond, corresponding partial structure is designed according to the type of minutiae. By syncretizing the local similarity of those partial structures, we can obtain the similarity of the whole image to judge whether the matching fingerprints come from the same finger. The experimental results indicate that this algorithm has certain advantage in the matching with small area.The quality of fingerprint image is one of the critical factors in fingerprint matching. The evaluation of quality is introduced as a key step in fingerprint matching. It can be classified into overall quality evaluation and local quality evaluation. This paper mainly focuses on the local quality evaluation of interesting region. Firstly, according to the type of each candidate minutea (i.e. the ridge ending and the ridge bifurcation), two local structures are constructed,as designed in Chapter 2: local structure of ending and local structure of bifurcation. Then, we get enclosing rectangle of each local structure, and evaluate the quality of each rectangle region to obtain quality index. Finally, we calculate the average of quality index of each minutia matched, and set the average factors as weights of similar minutia's local structure in overall fingerprint matching. The experimental results demonstrate that this method has improved the accuracy of fingerprint matching at some degree.