Index was outside the bounds of the array. 文章摘要
|本期目录/Table of Contents|

[1]张笑楠,张自友. 基于残差网络的三维人脸识别方法[J].内江师范学院学报(自然科学),2019,06:61-67.
点击复制

 基于残差网络的三维人脸识别方法(PDF)

《内江师范学院学报(自然科学)》[ISSN:1671-1785/CN:51-1521/Z]

期数:
2019年06期
页码:
61-67
栏目:
出版日期:
2019-06-25

文章信息/Info

Title:
-
文章编号:
1671-1785(2019)06-0061-07
作者:
 张笑楠12 张自友3
1.四川大学视觉合成图形图像技术国防重点学科实验室,
2.四川川大智胜软件股份有限公司
3.乐山师范学院物理与电子工程学院
Author(s):
-
关键词:
 三维人脸识别残差网络DLIBFRGCv2.0
Keywords:
-
分类号:
TP391.41
DOI:
10.13603/j.cnki.51-1621/z.2019.06.011
文献标识码:
A
摘要:
 针对目前三维人脸识别在时间成本和识别精度的平衡问题上,提出了一种基于残差网络的三维人脸
识别方法.该方法首先定义了一个二维平均人脸特征点和一个三维平均人脸特征点,将三维点云向三维平均人脸
特征点对齐后,做统一的透视投影得到深度图像,再经过人脸区域裁剪得到用于训练的数据,最后使用27层的残
差网络训练分类模型,从而实现三维人脸识别.由于提前设计平均人脸特征点,故大幅缩短了数据预处理时间,在
FRGCv2.0数据集上进行测试取得了很好的效果:中性对中性实验、对全部实验、对非中性实验,识别率分别为
98.8%、98.5%、98.5%,且总耗时仅为0.5秒.
Abstract:

参考文献/References

[1]郭丽丽,丁世飞.深度学习研究进展[J].计算机科学,2015,42(5):28-33.
[2]CREUSOTC ,PEARSN ,AUSTINJ.A Machine-Learning Approachto Keypoint Detectionand Landmarkingon3DMeshes [J].InternationalJournalofComputerVision,2013,102(1/3):146-179.
[3]NIXONK ,AIMALEV ,ROWER .SpoofDetectionSchemes[M].TheUnitedStates:HandbookofBiometricsSpringer,2008:403-423.
[4]KOLLREIDER K ,FRONTHALER H ,BIGUNJ .VerifyingLivenessbyMultipleExpertsinFaceBiometrics[C]//2008IEEEComputerSocietyConferenceonComputerVisionandPatternRecognition Workshops.IEEE,2008:1-6.
[5]徐俊,达飞鹏.一种基于局部描述符的三维人脸识别方法[J].模式识别与人工智能,2012,25(1):45-53.
[6]ZHONGC,SUNZ,TANT .Robust3DFaceRecognitionUsing Learned VisualCodebook [C]//IEEEConferenceonComputerVision& PatternRecognition.IEEE,2007:12-17.
[7]CHAU?aCQ,SILVAL ,BELLON ORP ,etal.3DFaceRecognition UsingSimulatedAnnealing AndTheSurfaceInterpenetration Measure.[J].IEEETransactionson Pattern Analysis & Machine Intelligence,2010,32(2):206-19.
[8]王跃明,潘纲,吴朝晖.三维人脸识别研究综述[J].计算机辅助设计与图形学学报,2008,20(7):819-829.
[9]CHUACS,HANF,HOYK .3DHumanFaceRecognitionUsingPointSignature[C]//InternationalConferenceonAutomatic Face & Gesture Recognition.Grenoble2000:233-238.
[10]LEE Y ,SONG H ,YANG U ,etal.LocalFeatureBased3DFaceRecognition[C]//InternationalConferenceonAudio-&Video-BasedBiometricPersonAuthentication.SpringerBerlinHeidelberg,2005:909-918.
[11]BERRETTIS,BIMBOAD,PALAP.3DFaceRecognitionUsingIsogeodesicStripes[J].IEEETransPatternAnalMachIntell,2010,32(12):2162-2177.
[12]KINGDE .Dlib-ml:A MachineLearningToolkit [J].JournalofMachineLearningResearch,2009,10(3):1755-1758.
[13]IOFFES,SZEGEDYC.BatchNormalization:AcceleratingDeepNetworkTrainingbyReducingInternalCovariateShift[C]//InternationalConferenceonInternationalConferenceonMachineLearning.JMLR.org,2015:448-456.
[14]SIMONYAN K,ZISSERMAN A.VeryDeepConvolutionalNetworksforLarge-ScaleImageRecognition [J].ComputerScience,2014(5):1-14.·66·第6期张笑楠,张自友:基于残差网络的三维人脸识别方法
[15]HE K,ZHANG X,REN S,etal.Deep ResidualLearningforImageRecognition[C]//ProceedingsoftheIEEEConferenceonComputerVisionandPatternRecognition.2016:770-778.
[16]WENY,ZHANGK,LIZ,etal.ADiscriminativeFeatureLearningApproachforDeepFaceRecognition [J].Computer& OperationsResearch,2016,47(9):11-26.
[17]LIH,HUANG D,MORVANJM,etal.Towards3DFaceRecognitionintheReal:A Registration-FreeApproach Using Fine-Grained Matchingof3D KeypointDescriptors[J].InternationalJournalofComputerVision,2015,113(2):128-142.
[18]Al-OSAIMIF ,BENNAMOUN M ,MIAN A .AnExpressionDeformationApproachtoNon-rigid3DFaceRecognition [J].International Journal of ComputerVision,2009,81(3):302-316.
[19]MAURER T ,GUIGONISD ,MASLOVI,etal.PerformanceofGeometrixActiveIDTM3DFaceRecognitionEngineontheFRGCData[J].InProceedingsofIEEE ConferenceonComputerVisionandPatternRecognition(CVPR),2005(3):145-154.
[20]WANGY,LIUJ,TANGX.Robust3DFaceRecognitionByLocalShapeDifferenceBoosting[J].IEEETransactionsonPattern Analysis & MachineIntelligence,2010,32(10):1858-1870.
[21]LEIY ,BENNAMOUN M ,HAYAT M ,etal.AnEfficient3D FaceRecognition Approach UsingLocalGeometrical Signatures [J]. Pattern Recognition,2014,47(2):509-524.

备注/Memo

备注/Memo:
更新日期/Last Update: 2019-07-04