基于频域统计量的联指图噪声检测和去除  

Slap fingerprint noise detection and removal algorithm based on frequency domain statistics

作　　者：李鹏[1] 张永良[1] 李焱淼[2] 李骏康[3] 

机构地区：[1]浙江工业大学计算机科学与技术学院,杭州310023 [2]大连交通大学理学院,辽宁大连116028 [3]浙江工业大学信息工程学院,杭州310023

出　　处：《计算机应用》2011年第A01期67-69,72,共4页journal of Computer Applications

基　　金：浙江省教育厅科研资助项目(Y201018424)

摘　　要：针对联指图中存在的各种噪声,提出了一种基于频域统计量的联指图噪声检测和去除算法。首先对联指图进行规格化处理,然后利用三个频域统计量来检测和去除噪声,最后对图像进行后处理。该算法根据指纹和各种噪声在频域上的不同特性,给出了相应的噪声判定条件。实验结果表明该算法是一种行之有效的联指图噪声检测和去除算法。To detect and remove the noise in slap fingerprint images,an efficient algorithm was presented based on frequency domain statistics.Firstly,the slap was normalized;then,three frequency domain statistics were used for noise detection and removal;finally,post-processing was applied to optimize the result.Based on the different frequency domain characteristics between fingerprint and noise,the corresponding criterion was proposed to detect and remove the noise.The experimental results show that the presented algorithm is effective.

关 键 词：指纹 联指图 噪声 频域统计量 规格化 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]