Fingerprint Image Segmentation And Matching Algorithm Research

In recent years, automatic fingerprint identification technique has been a research focus in the area of science. Many researchers have done much work on fingerprint identification. But some technical difficulties always baffle the development of fingerprint identification. The valid fingerprint segmentation algorithm and reliable, strong fingerprint-matching algorithm have important meaning to improving the speed and the performance of system.In this paper, we deeply analyze and investigate two key algorithms of automatic fingerprint identification system----fingerprint segmentation algorithm and minutiae matching algorithm. For fingerprint segmentation, aiming at the deficiency of existing method, we first bring forward a new thought of fingerprint segmentation based quadric surface model and carry it out. The experiments on representative databases indicate the validity of it. For fingerprint matching, do some improvement on the fingerprint minutiae-matching algorithm based on the local and global structures.The thesis is organized as follows: Chapter 1 is introduction, which introduces the general institutions and difficulties of fingerprint identification. Being the base of the studying, chapter 2 presents a fingerprint segmentation based on Block. Chapter 3 introduces a new fingerprint segmentation method based on quadric surface model, which is the main innovation of this article. Chapter 4 investigates one of fingerprint matching methods. Chapter 5 is conclusion and discussion. The main contents are as follows:Implement a fingerprint segmentation method based on block, which is the most popular algorithm in the field of Automatic Fingerprint Identification System. This method has the fast speed and it's easy to carry out. Experiments show that the segmentation only can reach block-wise, which is to say that the method decides which block is foreground or background, not the pixel. The border of foreground and background obtained by this method is rather serrate, which makes it difficult to judge the reliability of features extracted from the border.Aiming at the deficiency of the current fingerprint segmentation method at pixel and the segment limitation of the linear classifier, a new method of segmentation based on quadric surface model is proposed. We use three features of pixel, which are coherence, mean and variance, analyze the spatial distribution of pixels in foregroundand background, and then address a new method based on quadric surface model. This method can get the smooth border based on pixel. Experiments have shown that the proposed segmentation method has good performance and the segmentation result is obviously better than that of the linear classifier.A new fingerprint minutiae matching method based on the local and global structure is studied. Then aiming at the neighborhood selection deficiency of minutiae, we do some improvement on the fingerprint minutiae-matching algorithm based on the local and global structures. Experiment shows that this method has good performance in reliability, accuracy and processing speed.