%0 Journal Article %T 基于特征融合的卷积神经网络乳腺癌图像分类 %A 董永峰 %A 刘 %A 霞 %A 王利琴 %A 石陆魁 %J 河北工业大学学报 %D 2018 %R DOI:10.14081/j.cnki.hgdxb.2018.06.011 %X 传统乳腺癌图像分类方法需要从医学图像中人工提取特征,不仅需要具备专业医学知识,而且存在耗 时费力、提取高质量特征困难等问题.因此,提出了一种基于特征融合的卷积神经网络乳腺癌图像分类方法. 首先预训练了两个不同结构的卷积神经网络,然后利用卷积神经网络自动提取特征的特性,将两个结构提取 到的特征进行融合,最后利用分类器对融合的特征进行分类;同时,为避免卷积神经网络模型受小样本量限 制出现过拟合现象,通过乳腺病变区域提取、区域细化和数据增强等方法对图像进行适当预处理,并通过过 采样方法解决了正负样本不平衡的问题.实验结果显示,该方法在乳腺癌图像数据集BCDR-F03上分类AUC达 到89%,对乳腺癌图像的分类精度较传统方法有明显提高.</br>Traditionalbreastcancerimageclassificationmethodsneedtoextractfeaturesfrommedicalimages,which notonlyrequiresprofessionalmedicalknowledge,butalsoistime-consuming,laboriousanddifficulttoextracthighqualityfeatures.Therefore,aconvolutionneuralnetworkbasedonfeaturefusionisproposedforbreastcancerimageclas? sification.Thefirstpre-trainingoftwodifferentstructureoftheconvolutionalneuralnetwork,andthenusetheconvolu? tionneuralnetworkcharacteristicsofautomaticfeatureextraction,featureextractiontothetwostructuralintegration,fi? nallyusingclassifiertoclassifythefeatureoffusion;atthesametime,inordertoavoidtheconvolutionalneuralnetwork modelbythesmallsamplesizelimitoverfittingthisphenomenon,throughregionalextraction,refinementandenhance? mentofbreastlesionsofregionaldataandothermethodsofproperimagepretreatment,andthesamplingmethodtosolve theproblemoftheimbalanceofpositiveandnegativesamples.Theexperimentalresultsshowthatthismethodcanclassi? fyAUCinto89%onthedatasetBCDR-F03,andtheclassificationaccuracyofbreastcancerimagesissignificantlyim? provedcomparedwiththetraditionalmethods. %K 卷积神经网络 %K 特征融合 %K 支持向量机 %K 数据增强 %K 过采样 %K 乳腺癌病理图像< %K /br> %K convolutionalneuralnetwork %K featurefusion %K supportvectormachines %K dataaugmentation %K oversampling %K breastcancerhistologicalimage %U http://zrxuebao.hebut.edu.cn//oa/darticle.aspx?type=view&id=201806011