资源描述：
《交叉皮层模型及其在图像处理中的应用》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/290299405ImageprocessingusingintersectingcorticalmodelArticle·August2009CITATIONSREADS3215authors,including:KunZhanHongjuanZhangLanzhouUniversityLanzhouUniversity29PUBLICATIONS270CITATIONS13PUBLICAT
2、IONS131CITATIONSSEEPROFILESEEPROFILEYideMaLanzhouUniversity245PUBLICATIONS1,952CITATIONSSEEPROFILESomeoftheauthorsofthispublicationarealsoworkingontheserelatedprojects:ImageDetectionMethodsofBreastCancerinMammogramsViewprojectneuralnetworksprinciple&applicationViewprojectAllcontentfollowingthispag
3、ewasuploadedbyYideMaon24June2016.Theuserhasrequestedenhancementofthedownloadedfile.2009年8月北京邮电大学学报Aug.2009第32卷第4期JournalofBeijingUniversityofPostsandTelecommunicationsVol.32No.4文章编号:1007-5321(2009)04-0040-06交叉皮层模型及其在图像处理中的应用绽琨,张红娟,马义德,刘丽,田乐(兰州大学信息科学与工程学院,兰州730000)摘要:深入分析了交叉皮层模型(ICM)的性能参数、基本特性和工作原理
4、,得出无耦合时的内部活动项累加式、点火时刻表达式和点火周期等,总结出ICM模型表现的变阈值特性、非线性脉冲调制特性、同步脉冲发放现象、捕获特性、动态脉冲发放现象、自动波特性和综合时空特性.在此研究基础上提出自动图像分割算法和自动边缘提取算法以及用ICM与正交变换结合进行特征提取的算法,并采用ICM与数学形态学和中值滤波相结合的方法去除脉冲噪声.计算机仿真结果表明,提出的算法均能取得较好的结果,而且由于ICM结构简单,易于实现,运算速度也较快.关键词:交叉皮层模型;图像处理;分割;特征提取;平滑中图分类号:TP183;TP391文献标识码:AImageProcessingUsingInter
5、sectingCorticalModelZHANKun,ZHANGHong-juan,MAYi-de,LIULi,TIANLe(SchoolofInformationScienceandEngineering,LanzhouUniversity,Lanzhou730000,China)Abstract:Theworkingprinciple,performanceparameters,behaviorsandmaincharacteristicsofinter-sectingcorticalmodel(ICM)areanalyzed.Theexpressionsofinternalacti
6、vitywithoutcoupling,fir-ingtimeandfiringperiodicityarederivedfromthemathematicaldescriptionofICM.TheICMfea-turesofvariablethreshold,nonlinearpulsemodulation,synchronouspulsebursts,pulsecapture,dy-namicpulsebursts,auto-waveandsynthesizedspace-timearesummarized.Then,theapplicationsofICMinautomaticim
7、agesegmentation,automaticedgeextraction,featureextractionandimagesmoothingareintroducedrespectively.ExperimentsshowthatICMcanbeusedtosegmentimages,retrieveimagesandde-noiseimageseffectively.Keywords:intersectingc
