第5期 孙正兴，等：基于局部SM分类器的表情识别方法 ·465- ented a new LSVM classifier for classifying the expres- 2000.31:131-146 sions Experments on laboratory data Cohen-Ka- [9 ]L ISETTIC L,SCH ANO D J.Autmatic facial expression nade)showed 89 11%recogniton accuracy on 480 interpretation:where human-computer interaction,artificial video sequences At the same tme,we compared intelligence and cognitive science intersect[J].Pragatics &Cogniton,2000,8:185-235 KNN,SVM,and KNN-SVM classifiers with the LS- [10]ZHANG H,BERG A C,MA RE M,MAL IK J.SVM- VM.The LSVM classifier produced the best experi- KNN:Discrm inative nearest neighbor classificaton for mental results visual categry recognition[C]//Proceeding ofConference References: on Computer Vision and Pattem Recognition CVPR 2006).New York,USA:IEEE Computer Society,2006: [1 FRANCO L,TREVES A.A neural netork facial expres- 2126-2136 sion recogniton system using unsupervised bcal processing 11 ]BASSLI J.Emotion recognition:the role of facial move- [C]//Proceeding of the 2nd Intemational Symposium on ment and the relative mportance of upper and lower areas mage and Signal Processing and Analysis (SPA2001). of the face [J ]J Personality Social Psychol,1979,37: Pula,Croatia,2001:628-632 2049-2059 [2 ]L YONS M,AKAMATSU S Coding facial exp ressions with [12 ]EKMAN P,FR IESEN W V.Facial action coding system: gabor wavelets[C]//Proceeding of the Third IEEE Intema- investigator's guide [M ]Pab Alto:Consulting Psycholo- tional Conference on Automatic Face and Gesture Recogni- gists Press,1978:156-163. tion Nara,Japan:IEEE Computer Society,1998:454-459 [13 ]PANTC M,ROTHKRANTZL J M.Autmatic analysis of [3]BASSLI J.Eotion recognition:the ole of facial move- facial expressions the state of the art[J ]IEEE Trans on ment and the relative iportance of upper and bwer areas of PAML2000,22(12):1424-1445 the face[J]J Personality Social Psychol,1979,37:2049- [14 MASE K Recognition of facial expression from optical flw 2059 [J].IECE Transactons on Communications (Special Is- [4 FASEL B,LUETTN J.Recogniton of asymmetric facial sue on Computer Vision and its Applicatons),1991,10: action unit activities and intensities [C]//Poceeding of 3474-3483 15 th Intemational Conference on Pattem Recognition CPR [15 ]BLACKM J,YACOOB Y.Tracking and recognizing rigid 2000).Barcelona,Spain,2000:1100-1103. and non-rigid facial motions using bcal parametric models [5]TAN YL,KANADE T,COHN J F Recognizing action u- of mage motion [C]//Proceeding of Fifth Intemational nits or facial expression analysis[J].EEE Transacton on Conference on Computer Vision (CCV95).Cambridge, PAML2001,23(2):97-115 MA,USA,1995:374-381 [6]BARILETTM S,L ITILEWORT G,FRANKM G,LA N- [16]OL NER N,PENTLAND A,ERARD F B.LAFTER:a SCSEK C,FASEL I MOVELLAN J.Recognizing facial ex- real-time face and lips tracker with facial expression recog- pression:Machine leaming and applicaton o pontaneous nition [J ]Pattem Recognition,2000,33:1369-1382 behavior[C]//Proceeding of Conference on Computer Vi [17]BARTLETTM S,L ITILEWORT G.FASEL L MOVEL- sion and Pattem Recognition (CVPR 2005).San Dieg, LAN J R Real time face detection and facial expression CA,USA:IEEE Computer Society,2005:568-573. recognition:devebpment and applications to human com- [7]COHEN L SEVE N,COMMAN GG,CRELO M C, puter interaction[C]//Proceeding of Conference on Com- HUANG T S Leaming Bayesian netork classifier or facial puter Visin and Pattem Recognition (CVPR 2003). exp ression recognition using both labeled and unlabeled data Madison,Wisconsin,USA:IEEE Computer Society, [C]//Proceeding of Conference on Computer Vision and 2003:53-58 Pattem Recognition (CVPR 2003).Madison,W isconsin, [18]COHEN L SEBE N,GARG S,CHEN L S,HUANGA T USA:IEEE Computer Society,2003:595-601. S Facial expression recognition from video sequences [8 ]L IEN JJ,KANADE T,COHN J F,LIC C Detection, temporal and static modeling [J]Computer Vision and tracking,and classification of action units in facial expres mage Understanding.2003,91(1-2):160-187. sion [J].Joumal of Robotics and Autonomous Systems, [19]COOTES T F,TAYLOR C J,COOPER D H,GRAHAM 1994-2009 China Academic Journal Electronic Publishing House.All rights reserved.http://www.cnki.netented a new LSVM classifier for classifying the exp res2 sions. Experiments on laboratory data ( Cohen2Ka2 nade) showed 89. 11% recognition accuracy on 480 video sequences. A t the same time, we compared KNN, SVM, and KNN2SVM classifiers with the LS2 VM. The LSVM classifier p roduced the best experi2 mental results. References: [ 1 ] FRANCO L, TREVES A. A neural network facial exp res2 sion recognition system using unsupervised local p rocessing [C ] / / Proceeding of the 2nd International Symposium on Image and Signal Processing and Analysis ( ISPA2001 ). Pula, Croatia, 2001: 6282632. [ 2 ]LYONS M, AKAMATSU S. Coding facial exp ressions with gabor wavelets[C ] / /Proceeding of the Third IEEE Interna2 tional Conference on Automatic Face and Gesture Recogni2 tion. Nara, Japan: IEEE Computer Society, 1998: 4542459. [ 3 ]BASSIL I J. Emotion recognition: the role of facial move2 ment and the relative importance of upper and lower areas of the face[J ]. J Personality Social Psychol, 1979, 37: 20492 2059. [ 4 ] FASEL B, LUETTIN J. Recognition of asymmetric facial action unit activities and intensities [ C ] / /Proceeding of 15 th International Conference on Pattern Recognition ( ICPR 2000). Barcelona, Spain, 2000: 110021103. [ 5 ] TIAN Y L, KANADE T, COHN J F. Recognizing action u2 nits for facial exp ression analysis[J ]. IEEE Transaction on PAM I, 2001, 23 (2) : 972115. [ 6 ]BARTLETTM S, L ITTLEWORT G, FRANK M G, LA IN2 SCSEK C, FASEL I, MOVELLAN J. Recognizing facial ex2 p ression: Machine learning and app lication to spontaneous behavior[C ] / / Proceeding of Conference on Computer V i2 sion and Pattern Recognition (CVPR 2005 ). San D iego, CA, USA: IEEE Computer Society, 2005: 5682573. [ 7 ] COHEN I, SEVE N, COZMAN G G, CIRELO M C, HUANG T S. LearningBayesian network classifier for facial exp ression recognition using both labeled and unlabeled data [C ] / / Proceeding of Conference on Computer V ision and Pattern Recognition (CVPR 2003). Madison, W isconsin, USA: IEEE Computer Society, 2003: 5952601. [ 8 ]L IEN J J, KANADE T, COHN J F, L I C C. Detection, tracking, and classification of action units in facial exp res2 sion [ J ]. Journal of Robotics and Autonomous Systems, 2000, 31: 1312146. [ 9 ]L ISETTI C L, SCH IANO D J. Automatic facial exp ression interp retation: where human2computer interaction, artificial intelligence and cognitive science intersect[J ]. Pragmatics & Cognition, 2000, 8: 1852235. [ 10 ] ZHANG H, BERG A C, MA IRE M, MAL IK J. SVM2 KNN: D iscriminative nearest neighbor classification for visual category recognition[C ] / / Proceeding of Conference on Computer V ision and Pattern Recognition ( CVPR 2006). New York, USA: IEEE Computer Society, 2006: 212622136. [ 11 ]BASSIL I J. Emotion recognition: the role of facial move2 ment and the relative importance of upper and lower areas of the face [ J ]. J Personality Social Psychol, 1979, 37: 204922059. [ 12 ] EKMAN P, FR IESEN W V. Facial action coding system: investigator’s guide [M ]. Palo A lto: Consulting Psycholo2 gists Press, 1978: 1562163. [ 13 ] PANTIC M, ROTHKRANTZ L J M. Automatic analysis of facial exp ressions: the state of the art[J ]. IEEE Trans on PAM I, 2000, 22 (12) : 142421445. [ 14 ]MASE K. Recognition of facial exp ression from op tical flow [J ]. IEICE Transactions on Communications ( Special Is2 sue on Computer V ision and its App lications) , 1991, 10: 347423483. [ 15 ]BLACK M J, YACOOB Y. Tracking and recognizing rigid and non2rigid facial motions using local parametric models of image motion [ C ] / /Proceeding of Fifth International Conference on Computer V ision ( ICCV95 ). Cambridge, MA, USA, 1995: 3742381. [ 16 ]OL IVER N, PENTLAND A, ERARD F B. LAFTER: a real2time face and lip s trackerwith facial exp ression recog2 nition[J ]. Pattern Recognition, 2000, 33: 136921382. [ 17 ]BARTLETT M S, L ITTLEWORT G, FASEL I, MOVEL2 LAN J R. Real time face detection and facial exp ression recognition: development and app lications to human com2 puter interaction[ C ] / /Proceeding of Conference on Com2 puter V ision and Pattern Recognition ( CVPR 2003 ). Madison, W isconsin, USA: IEEE Computer Society, 2003: 53258. [ 18 ]COHEN I, SEBE N, GARG S, CHEN L S, HUANGA T S. Facial exp ression recognition from video sequences: temporal and static modeling [ J ]. Computer V ision and Image Understanding, 2003, 91 (122) : 1602187. [ 19 ]COOTES T F, TAYLOR C J, COOPER D H, GRAHAM 第 5期 孙正兴 ,等 :基于局部 SVM分类器的表情识别方法 ·465·