基于小波域软阈值的指纹分割方法研究  

A fingerprint image segmentation based on soft-threshold wavelet domain

作　　者：肖刚[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,哈尔滨150001

出　　处：《哈尔滨商业大学学报（自然科学版）》2012年第2期231-234,共4页Journal of Harbin University of Commerce：Natural Sciences Edition

摘　　要：鉴于指纹图像的质量对于其后续识别步骤具有非常重要的影响,因而需要对采集的指纹图像进行增强处理.提出了一种新的基于小波域软阈值Contourlet变换的指纹图像分割算法.Contourlet变换是一种新的多尺度几何分析方法,它不仅拥有小波变换的时频局域特性和多分辨率特性,而且还具有很好的方向性和各向异性.实验结果表明,该算法能够有效提升指纹图像的质量,提高指纹纹线谷脊清晰度,且具有较好的连续性.The quality of fingerprint image has great affect for its follow-up steps such as recognition. Therefore, an effective method is needed to enhance the collected fingerprint image. In this paper, a new fingerprint image enhancement approach based on Contourlet transform soft-threshold wavelet domain was proposed. Contourlet is a new multi-scale geo- metric analysis method, which not only has the main features of local time-frequency and multi-resolution, but also has the directionality and anisotropy. Experimental results demon- strated that the method could effectively improve the quality of fingerprint image, and could increase the clarity of fingerprint ridges and valleys, and have better continuity.

关 键 词：指纹分割 模极大值 轮廓波 CONTOURLET变换 方向滤波 

分 类 号：TN911.73[电子电信—通信与信息系统]