Minutiae-Based Research In Fingerprint Recognition

With modern electronic technology developing so fast, computer systems security becomes an important issue. Traditional security systems have been insufficient to meet the needs, and therefore, we are doing research on biometric technology. Fingerprint Idenification, one of the most representative technologies, has made tremendous advance and been applied to various electronic equipment by many companies and organizations. However, as automated fingerprint identification technology involves commercial property, many algorithms are unable to open source, so the fingerprint Idenification technology is still one of researches focused nowadays.In this paper, we study the principle of fingerprint Idenification based on fingerprint identification and image processing algorithms firstly. This paper mainly includes the following areas:In the fingerprint image preprocessing, image smoothing techniques are used with basic filter to delete the noise of images in collected information. Furthermore, we put forward a modified Otsu threshold segmentation method, which can segment the fingerprint image and the background image.As the directional specialty of ridge, we discuss the Gabor principle in this paper and construct a directional filter to enhance the fingerprint image, which can improve the clarity of ridge and reduce the impact of noise on the feature extraction. At the same time, we use the existing image thinning algorithm to thin fingerprint ridge to facilitate the extraction of feature points.On fingerprint feature extraction and matching, this paper presents a model based on minutiae features fingerprint Idenification algorithm based on the thinned image, using eight neighborhood template to extract thinned fingerprint image endpoint and bifurcation points, and collected the false feature points. At last, with details of the points collections, we complete match the collected data points.All in all, we reduce the noise from the above-mentioned study and enhance the fingerprint details. Also we give the specific solutions and results about feature extraction and matching the model. The experimental results contribute to automatic fingerprint Idenification technology and the practical application.