指纹图像分割算法的分析与实现

《指纹图像分割算法的分析与实现》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、山东师范大学硕士学位论文3.最后本文在MatlabR2008环境中进行了仿真实验，分别比较了基于CMVF特征的指纹图像分割算法、基于Adaboost分类器结合CMV特征以及基于Adaboost分类器结合CMVF特征分割算法的实验结果。数据显示，采用CMVF特征进行分割的效果要比仅仅使用CMV特征要精确很多。因此可以得出，对指纹图像描述越详细的特征越有利于指纹图像的分割，但是在添加特征时，一定要注意防止特征的冗余，避免过大的增加算法的时间复杂度。关键词：自动指纹识别系统，指纹图像分割，Adaboost，CMV，Otsu中图分类号：TP391

2、ii万方数据山东师范大学硕士学位论文ResearchandRealizeofSegmentationAlgorithmofFingerprintImageABSTRACTInthefieldofbiologicalrecognitionsystem,now,automaticfingerprintidentificationsystem(AFIS)hasbeenwidelyacceptedandappliedalthoughtheapplicationoffingerprintidentificationsystemhasbeenvery

3、matureandpeoplealsohaveagoodunderstandingofeachstepoftherecognitionsystem,butwiththedevelopmentofscienceandtechnologyandthesocialproblems,thestudyoffingerprintrecognitionneeddeeperresearchtosolvetheemergingproblemsintheapplicationoffingerprintidentification.AFIScanbedivid

4、edintofoursteps:theacquisitionoffingerprintimage;preprocessingofthefingerprintimage;extractfeature,fingerprintmatching.Inordertoreducetheredundancyofunnecessarycomputation,fingerprintsegmentationneedsinfrontoffingerprintimagepreprocessing.Thisarticlewillfocusonthefingerpr

5、intimagesegmentationtechniquesandalgorithmstodomorein-depthstudyandmakeasimulationexperimentoneachsegmentationalgorithm.Fingerprintimagesegmentationisanimportantpartofthepreprocessingoffingerprintandtheautomaticfingerprintrecognitionsystem,Itcangreatlyreducesubsequentproc

6、essingcomputationandsubsequentprocessingtimetoseparatedpreciselyfingerprintforegroundareafromthebackgroundregionandwillsignificantlyincreasetheprocessingspeedandqualityofautomaticfingerprintidentificationsystem.Themainresearchcontentsofthispaperareasfollows:1.First,Thispa

7、perpresentsakindofsegmentationalgorithmoffingerprintimagebasedonCMVFfeatures.becauseTheclassicalalgorithmofsinglefeatureorsimplefusionaredifficulttosegmentfigerprintimageeffectlyforlowqualityfingerprintimage,sothesegmentationalgorithmoffingerprintimagebasedonCMVFfeaturesi

8、sproposed.Thisalgorithmusesthreeclassicfeatures,therearethefingerprintdirectionconsistency,avera