一种新型指纹图像分割算法

作者：黄 斌 日期：2010-09-08/span> 浏览：3543 查看PDF文档

一种新型指纹图像分割算法

黄 斌

（浙江工业大学 信息工程学院，浙江 杭州 310014）

摘要：针对传统分割算法在含污迹较多的指纹识别应用中分割效果较差的问题，首先介绍了指纹图像分割的基本概念；然后对分割算法在输入指纹图像携带一定污迹、杂质等噪声环境下的鲁棒性进行了重点研究；最后，提出了一种新的基于最大类间方差法和传统方差法的分块处理的分割算法，并通过自适应滤波进一步平滑指纹图像的脊线及边界轮廓。实验结果表明，该方法较传统的分割算法能更好地完成实际应用中含污迹杂质较多的指纹的分割处理。
关键词：指纹识别；图像分割；图像预处理
中图分类号：TP39
文献标识码：A文章编号：1001-4551(2010)07-0078-03

A new segmenting algorithm for fingerprint images
HUANG Bin
(College of Imformation Engineering, Zhejiang University of Technology, Hangzhou 310014, China)

Abstract: Aiming at the problem that the traditional segmenting algorithms can not be effectively used in the fingerprint identification while the finger contain many impurities, firstly, the basic concepts of segmenting algorithms were introduced. The robustness of the segmenting algorithms to the noises attached to the input fingerprint images, such as dirt, impurities and low contrast was analyzed. A new method to segment the fingerprint images was proposed based on the maximum betweenclass variance and subblock processing of the traditional variance segmenting. And the adaptive filtering was used to smooth the ridge and boundary contours of fingerprint images in this method. The test results indicate that, compared with the traditional segmenting algorithms, the proposed method can segment better the fingerprint images that contain many impurities and stains in practice.
Key words: fingerprint identification; image segmentation; image preprocessing