一种改进小波阈值函数的指纹图像去噪方法

打开文本图片集

关键词：指纹图像；小波阈值去噪；阈值函数；峰值信噪比；均方根误差DOI：10.15938/j. jhust.2025.02.015中图分类号：TN911.73 文献标志码：A 文章编号：1007-2683（2025）02-0140-10

Abstract：Fingerprintrecognitiontechnologyisoftendisturbedbynoise.Intheprocessofusing waveletthresholddenoising methodtoreducefingerprintimagenoise，theselectionofthethresholdfunctionisparticularlysignificant.Therefore，anewthreshold functionwithadjustableparameters，whichisbetweentraditionalsoftandhardthresholdfunctions，isconstructed.Theewthresold functionovercomesthediscontiuityofthehardthresholdfunctionatthetresholdpoint，improvestheconstantdeviationoftheoft thresholdfunctionandcanreachzerodeviationundercertainconditions.Simulationexperimentsshowthatusingthenewthreshold functiontoremovesaltand-pepperoiseandGausianoiseifingerprintimagesduringteavelet tresholddnosingprocsults inlowerrot meansquareeroranddegreeof distorioncomparedtousingtraditionalandimprovedthreshold functions，while increasingpeaksigaltonoiseatiooelationcoeffcient，ndstrucuralsilarityindexmeasure.Tisotonlyprovidesodto improve thethresholdfunction，butalsoofersaneffctive meansforimprovingthedenoisingqualityoffingerprintimages.

Keywords：fingerprintimage；waveletthresholddenoising；thresholdfunction；peak signal to noiseratio；root mean square eror

0 引言

指纹是人们的个体特征，代表着每个人独特的身份信息。（剩余11196字）