Research On Fingerprint Feature Extraction Method Based On Grayscale Image

As the fingerprint identification technology develops, its applied fields get more and more widely. Nevertheless the algorithm of the fingerprint identification technology is complex, and hard to improve the speed of identification, which makes it limited in the application. As the real-time requirements continue to improve, people begin to realize that improving the speed of fingerprint feature extraction and reducing the complexity of algorithm is very important.Automatic fingerprint identification method commonly uses point mode to extract fingerprint characteristics. This method has good recognition accuracy, and the algorithm is complex, and time-consuming. Fingerprint feature extraction method mostly uses template form, whose speed improves obviously. But as the algorithm is complex, the speed is slow, when the demand is higher in real-time, it can't meet the practical demands. Not through complex processing but direct method of extracting fingerprint characteristics from gray image, improves the speed obviously. There are many documents that introduced many achievements on this research. But it selects many threshold values, and the requirement for the quality of fingerprints is high, so it's hard to be used.This article divides fingerprint characteristic extraction method based on the gray image into two steps, fingerprint image segmentation and feature extraction. It's mainly used to solve the speed problem. In fingerprint image segmentation, the article proposes fingerprint image segmentation method based on gauss operator. This method uses gauss operator to extract the boundary lines of fingerprint, and uses morphological method, divides up fingerprint image of background and fingerprints valid area availably. Gauss operator can extract fingerprint boundary availably, and has good filtering properties on lines of fingerprint, is favorable for fingerprint segmentation and subsequent processing. Experiments prove that this method has high speed and high precision, for the fingerprint image of poor quality also has very good segmentation effect. Extracting fingerprint feature points, this article puts forward method for extracting fingerprint characteristics based on the curvature of the testing. This method uses characteristics of fingerprint lines, curvature of endpoint and bifurcation point is larger. It detects curvature of fingerprint image, and extracts the points whose curvature is larger, so it can get characteristics of fingerprint image. Experiments prove that this method can extract fingerprint feature points effectively, and has higher speed.This article pays attention to effect of feature extraction and time of algorithm process, divides up parts of the image and extracts feature points, compares the results of process with traditional point model of the method for extracting fingerprint feature points. The results indicate that the method for extracting fingerprint feature used less time, so it shows that this method has very good advantage on speed.