摘要
指纹增强对于提高细节特征提取的准确率乃至整个自动指纹识别系统的性能实现都具有重要的意义.研究了基于加博(Gabor)函数的指纹增强算法,对纹线方向提取算法进行了改进,提出了一种纹线频率提取方法,给出了加博函数用于指纹增强的具体应用形式,以指纹图像的纹线方向和纹线频率为参数,使用加博函数实现了对指纹图像的增强处理,并使用南京大学活体指纹库中的部分典型低质量图像样本对算法性能进行了测试.实验结果表明,该算法对低质量图像具有显著的增强效果,增强后可以有效地降低指纹匹配的拒识率(false reject rate,简称FRR),处理速度和增强效果可以基本上满足在线自动指纹识别系统的需要.
Fingerprint enhancement is important for improving the accuracy of minutiae extraction, even for the total automated fingerprint identification system (AFIS). A fingerprint enhancement algorithm based on Gabor function and its application is developed in this paper. A method for ridge orientation acquirement is improved and a new method for ridge frequency acquirement is brought out. The formula that Gabor function is used in fingerprint enhancement is presented, and the fingerprint enhancement is implemented using ridge orientation and ridge frequency as parameters. The performance of this method is evaluated with some typical low quality images of Nanjing University fingerprint database. Experimental results indicate that the method has dramatic effect to low quality fingerprint images and false reject rate (FRR) of fingerprint matching can be effectively decreased after enhancement. Speed of the method can fill the demand of on-line AFIS.
出处
《软件学报》
EI
CSCD
北大核心
2003年第3期484-489,共6页
Journal of Software
基金
Supported by the Science and Technology Development Plan Program of Changchun City of China under Grant No.99011 (长春市科技发展计划)
the Application Development Foundation of Nanjing University of China under Grant No.2001-03 (南京大学应用开发基金)
关键词
加博函数
指纹增强算法
图像识别
指纹图像
图像处理
fingerprint
Gabor function
fingerprint enhancement
ridge orientation
ridge frequency