预览加载中,请您耐心等待几秒...
1/10
2/10
3/10
4/10
5/10
6/10
7/10
8/10
9/10
10/10

亲,该文档总共17页,到这已经超出免费预览范围,如果喜欢就直接下载吧~

如果您无法下载资料,请参考说明:

1、部分资料下载需要金币,请确保您的账户上有足够的金币

2、已购买过的文档,再次下载不重复扣费

3、资料包下载后请先用软件解压,在使用对应软件打开

一种基于面部纹理特征融合的人脸表情识别方法高婷婷李航殷守林摘要:针对人脸表情识别领域受噪声和遮挡等因素影响识别率不高的问题,结合局部和全局特征,提出一种基于面部表情的情感分析混合方法。首先,通过将梯度直方图(HOG)与复合局部三元模式(C-LTP)融合来进行特征提取;其次,将HOG和C-LTP提取的特征融合到单个特征向量中;最后,采用多类支持向量机分类器把特征向量进行情感分类;最后,将提出的方法在3个公共表情图像数据库中与现有的表情识别方法进行对比实验。结果表明,提出的方法在MMI,JAFFE,CK+数据库上的正确识别率分别为98.28%,95.75%,99.64%,平均识别率比其他方法高出10%,优于其他现有的方法。提出的表情识别方法,可有效促进人机交互系统的发展和计算机图像理解的研究,对实现人体语言与自然语言的融合,以及语言与表情连接模型的建立与实现具有重要意义。关键词:模式识别;人脸表情识别;特征融合;HOG;C-LTP;支持向量机:TP957.52:Adoi:10.7535/hbkd.2021yx02004AfacialexpressionrecognitionmethodbasedonfacetexturefeaturefusionGAOTingting,LIHang,YINShoulin(SoftwareCollege,ShenyangNormalUniversity,Shenyang,Liaoning110034,China)Abstract:Aimingatfacialexpressionrecognition,therecognitionrateisnothighduetonoiseandocclusion.Ahybridapproachoffacialexpressionhasbeenpresentedbycombininglocalandglobalfeatures.First,featureextractionisperformedtofusethehistogramoforientedgradients(HOG)descriptorwiththecompoundedlocalternarypattern(C-LTP)descriptor.Second,featuresextractedbyHOGandC-LTParefusedintoasinglefeaturevector.Third,thefeaturevectorissenttoamulti-classsupportvectormachineclassifierforfacialclassification.Finally,theproposedmethodiscomparedwiththeexistingfacialexpressionrecognitionmethodsinthreepublicfacialexpressionimagedatabases,andtheresultsshowthattherecognitionratesoftheproposedmethodinMMI,JAFFEandCK+databasesare98.28%,95.75%and99.64%,respectively.Theaveragerecognitionrateis10%higherthanothermethods,whichisbetterthanotherexistingmethods.Theresultsofthisstudyprovideareferencefortheresearchoffacialexpressionrecognitioninmanysituations.Themethodoffacialexpressionrecognitionproposedcaneffectivelypromotethedevelopmentofhuman-computerinteractionsystemandthestudyofcomputerimageunderstanding.Itisofgreatsignificancetorealizethefusionofhumanlanguageandnaturallanguage,aswellastheestablishmentandimplementationoftheconnectionmodelbetweenlanguageandexpression.Keywords:patternrecognition;facialexpressionrecognition;featurefusion;HOG;C-LTP;supportvectormachine面部表情[1]是人際关系中非常重要的交流方式。人脸表