基于混合高斯隐马尔科夫模型的运动对象轨迹识别
摘要
基于混合高斯隐马尔科夫模型(GHMM)的轨迹识别中,关键的问题是运动对象轨迹的有效获取,以及基于这些轨迹的健壮表示和建模,然后根据模型来识别轨迹。本文以视频运动对象轨迹识别为技术目标,从低层视频分析技术开始,系统研究了轨迹识别框架下的目标检测、跟踪和轨迹建模与识别算法,重点研究了轨迹建模与识别,采用混合高斯隐马尔科夫模型对获取的轨迹进行建模。为了降低特征空间的维数,减少计算量,把轨迹用参数函数的形式表示成一条曲线,在曲率变化最大的点处分割轨迹为一些相似的极小单元,称其为子轨迹,用它们的主成分分析(PCA)系数来表示该轨迹,用PCA系数来训练混合高斯隐马尔科夫模型,并且对运动目标轨迹进行识别。
本文分析视频来自CAVIAR项目,是从固定摄像头获取的,对顾客在超市门口活动的轨迹进行识别,从而得知顾客的具体活动。我们把顾客的活动分为六类:从左到右,从右到左,从左边进去,从右边进去,从左边出来和从右边出来。具体实现过程分为以下几步:首先实现场景中人的检测和跟踪,并提取轨迹;然后用均值聚类法对轨迹聚类,把轨迹分为六类,对每类轨迹进行PCA系数表示;最后对每类轨迹建立GHMM,通过计算每个轨迹的识别率来验证模型的有效性。实验结果表明,基于PCA系数表示的GHMM有着很好的轨迹识别能力。通过与整个轨迹点为输入向量的GHMM识别比较,该方法不仅降低了特征空间维数,还提高了识别率。
In the field of trajectory recognition based Gaussian Mixture Hidden Markov Model (GHMM), the key issues is obtaining trajectory of moving objects, robust presenting and modeling of the trajectory. Furthermore, we analyze activities by the models. Given technical objectives of trajectory recognition, this paper investigates systemically many Algorithms, following the framework of trajectory recognition, such as moving objective detecting, Tracking and trajectory modeling etc. We concentrated on trajectory modeling. We model trajectory by Gaussian Mixture Hidden Markov Model. Before modeling, trajectories are segmented at points of change in curvature, and the subtrajectories are represented by their principal component analysis (PCA) coefficients, to reduce the feature space dimension. Then we present a framework to robustly estimate the multivariate probability density function based on PCA coefficients of the subtrajectories using GHMMs.
The Experimental video is from CAVIAR project in the paper, which is obtained from a fixed camera. We can analysis activity of customers’coming and going by trajectory recognition. The customers’activities include walking from left to right, walking from right to left, entering from the left, entering from the right, out from the left and out from the right. This process is divided into three stages. First, we obtain the trajectory through detecting and tracking. Second, we cluster the trajectory into six categories by K-means clustering, and PCA reducing dimensionality is used to process trajectory; Finally, we will establish GHMM for each trajectory, and Validate the validity of the model by calculating the recognition rate of each trajectory. Experiments demonstrate the superiority of our proposed HMM-based scheme using PCA coefficients of subtrajectories in comparison with other techniques in the literature.
本文分析视频来自CAVIAR项目,是从固定摄像头获取的,对顾客在超市门口活动的轨迹进行识别,从而得知顾客的具体活动。我们把顾客的活动分为六类:从左到右,从右到左,从左边进去,从右边进去,从左边出来和从右边出来。具体实现过程分为以下几步:首先实现场景中人的检测和跟踪,并提取轨迹;然后用均值聚类法对轨迹聚类,把轨迹分为六类,对每类轨迹进行PCA系数表示;最后对每类轨迹建立GHMM,通过计算每个轨迹的识别率来验证模型的有效性。实验结果表明,基于PCA系数表示的GHMM有着很好的轨迹识别能力。通过与整个轨迹点为输入向量的GHMM识别比较,该方法不仅降低了特征空间维数,还提高了识别率。
In the field of trajectory recognition based Gaussian Mixture Hidden Markov Model (GHMM), the key issues is obtaining trajectory of moving objects, robust presenting and modeling of the trajectory. Furthermore, we analyze activities by the models. Given technical objectives of trajectory recognition, this paper investigates systemically many Algorithms, following the framework of trajectory recognition, such as moving objective detecting, Tracking and trajectory modeling etc. We concentrated on trajectory modeling. We model trajectory by Gaussian Mixture Hidden Markov Model. Before modeling, trajectories are segmented at points of change in curvature, and the subtrajectories are represented by their principal component analysis (PCA) coefficients, to reduce the feature space dimension. Then we present a framework to robustly estimate the multivariate probability density function based on PCA coefficients of the subtrajectories using GHMMs.
The Experimental video is from CAVIAR project in the paper, which is obtained from a fixed camera. We can analysis activity of customers’coming and going by trajectory recognition. The customers’activities include walking from left to right, walking from right to left, entering from the left, entering from the right, out from the left and out from the right. This process is divided into three stages. First, we obtain the trajectory through detecting and tracking. Second, we cluster the trajectory into six categories by K-means clustering, and PCA reducing dimensionality is used to process trajectory; Finally, we will establish GHMM for each trajectory, and Validate the validity of the model by calculating the recognition rate of each trajectory. Experiments demonstrate the superiority of our proposed HMM-based scheme using PCA coefficients of subtrajectories in comparison with other techniques in the literature.
引文
[1]Moeslund,T.B.,Hilton,A.,Krüger,V.:A survey of advances in vision-based human motion capture and analysis[J]. Computer Vision and Image Understanding. 2006, 104(2):90-126.
[2]王亮,胡卫明,谭铁牛.人运动的视觉分析综述[J].计算机学报, 2002,25(3):225-237.
[3]纪运红.智能监控系统若干关键技术的研究和实现[D].辽宁科技大学.2007.
[4]Buzan,D.,Sclaroff,S.,Kollios,G.:Extraction and clustering of motion trajectories in video[A]. In: Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, UK, 2004:521-524.
[5]Lee,K.K.,Yu,M.,Xu,Y.S.:Modeling of human walking trajectories for surveillance[C]. In: Proceedings of IEEE International Conference on Intelligent Robots and Systems, 2003, 2:1554-1559.
[6]Makris,D.,Ellis,T.:Learning semantic scene models from observing activity in visual surveill -ance[C].In:IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 2005, 35(3):397-408.
[7]胡卫明,谢丹,谭铁牛.轨迹分布模式学习的层次自组织神经网络方法[J].计算机学报,2003, 26(4):417-426.
[8]Johnson,N.,Hogg,D.:Learning the distribution of objects trajectories for event recognition[J].Image and Vision Computing, 1996, 14(8):609-615.
[9] Owens,J., Hunter,A.:Application of the self-organizing map to trajectory classification[C].In:Third IEEE International Workshop on Visual Surveillance, 2000, p77-83.
[10]Piciarelli,C.,Foresti,G.L.:Online trajectory clustering for anomalous events detection[J].Pattern Recog -nition Letters, 2006, 27(15):1835-1842.
[11]Li,X.,Hu,W.M.,Hu,W.:A coarse-to-fine strategy for vehicle motion trajectory clustering[A].In:Proceedings International Conference on Pattern Recognition, 2006, 1:591-594.
[12]Hu,W.,Xiao,X.,Xie,D.,Tan,T.and Maybank,S.:A system for learning statistical motion patterns[J]. IEEE Trans. Pattern Anal. Machine Intell, 2006, 28(9):1450-1464.
[13]Naftel,A. Khalid,S.:Classifying spatiotemporal object trajectories using unsupervised learning in the coefficient feature space[J]. ACM Multimedia Systems, 2006, 12(3):227-238.
[14]Yu,J.W.,Li,S.,Chen,Y.S:Trajectory-based video retrieval by string matching [C].In:Proceedings International Conference on Image Processing, ICIP, 2004,4:2243-2246.
[15]Stauffer,C.,Grimson,W.,Eric,L.:Learning patterns of activity using real-time tracking[J].Transactions onPattern Analysis and Machine Intelligence, IEEE, 2000, 22(8):747-757.
[16]Hu,W.M.,Xiao,X.J.,Xie,D.,Tan,T.N.,Maybank,Steve.:Traffic accident prediction using 3-D model-based vehicle tracking[J].Transactions on Vehicular Technology, IEEE, 2004, 53(3):677-694.
[17]Naftel,A.,Khalid,S.:Classifying spatiotemporal object trajectories using unsupervised learning in the coefficient feature space[J].Multimedia Systems, 2006, 12(3):227-238.
[18]Brand,M.,Oliver,N.,Pentland,A.:Coupled hidden Markov models for complex action recognition [C].In:Proc. of IEEE CS Conf. on Computer Vision and Pattern Recognition, 1997:994-999.
[19]Zhou,H.,Don,K.:Unusual event detection via multi-camera video mining[C].In:Proc. of the 18th International Conf. on Pattern Recognition, ICPR archive,2006, 3:1161-1166.
[20]李和平,胡占义,胡毅红,吴福朝.基于半监督学习的行为建模与异常检测[J].Jornal of Software,2007, 18(3):527-537.
[21]Vinod,N.,James,J.C.:Automated visual surverllance using hidden marknov models[J].Centre for Intelligent Machines, Mcgill University, Montreal, PQH3A2A7.2002.
[22]Naohiko,S.,Kosuke,H.,Kenichi,T.,Yoshinori,K.,Yoichi,S.:Learning motion patterns and anomaly detection by human trajectory analysis[C]. In: IEEE International Conference on Digital Object Identifie.2007: 498-503
[23]Morris,B.T.,Trivedi,M.M.:A survey of vision-based trajectory learning and analysis for surveillance [J].IEEE Transactions on Circuits and Systems for Video Technology, 2008, 18(8):1114-1127.
[24]Bashir,F.,Khokhara,S.D.:Automatic object trajectory-based motion recognition using gaussian mixture models[C].In:International Conference on Multimedia&Expt, IEEE 2005:1532-1535
[25]潘奇明,程咏梅.基于隐马尔科夫模型的运动目标轨迹识别[J].计算机应用研究,2008,25(7)
[26]张丽君,吴晓娟,盛赞.基于HMM复杂场景下的行为识别方法[J].2008,(7):07-076.
[27]于殿泓.图像检测与处理技术[M].西安电子科技大学出版社.2007.1:100-175.
[28]Verri,A.,Uars,S., DeMieheli,E.:Motion segmentation from optical flow[C].In: Proc. the 5th Alvey Vision Conf., Brighton, UK, 1989:209-214.
[29]Barron,J.,Fleet,D.,Beauehemin,S.:Performance of optical flow techniques[J].International Journal of Computer Vision, 1994:12(l):42-77.
[30]Tong,K.,Yun,J.H., et al .:Motion estimation using new blocks based on the frame difference for frame rate-up conversion[C].In:Proceedings of International Meeting on Information Display, 2008, 8:1043-1046.
[31]Zhan,C.H.,Duan,X.H.,Xu,S.Y.,et al.:An improved moving object detection algorithm based on frame difference and edge detection[C].In:Proceedings of the 4th International Conference on Image and Graphics, ICIG, 2007:519-523.
[32]Mahadevan,V.,Vasconcelos,N.:Background subtraction in highly dynamic scenes[C].In:26th IEEE Conference on Computer Vision and Pattern Recognition CVPR, 2008:1-6
[33]Thongkamwitoon,T.,Aramvith,S.,Chalidabhongse,T.H. :An adaptive real-time background subtraction and moving shadows detection[C].In:IEEE International Conference on Multimedia and Expo (ICME), 2004, 2:1459-1462.
[34]马义德,朱望飞,安世霞,邱会银,汤书森.改进的基于高斯混合模型的运动目标检测方法[J].计算机应用, 2007,27(10):2544-2546.
[35]Zivkovic,Z.:Improved adaptive Gaussian mixture model for background subtraction[A].In:Proceedings International Conference on Pattern Recognition, 2004, 2:28-31.
[36]Cheng,J.,Yang,J.,Zhou,Y.:A novel adaptive Gaussian mixture model for background subtraction[C].Pattern Recognition and Image Analysis: Second Iberian Conference, Lecture Notes in Computer Science, 2005, 3522(I):587-593.
[37]郝彦彬.基于视频序列图像的运动目标检测与跟踪[D].大连理工大学.2008.
[38]David,B.,Philip,M.,et al.:A Real-time Computer Vision System for Measuring Traffic Parameters[J]. Computer Vision and Pattern Recognition, IEEE, 1997:495-501
[39]Huang,F.Z.,Su,J.B.:Abstracting and tracking of outline of face based on level set method[J].Computer transaction, 2003, 26(4):491-496
[40]Rucklidge W. J.:Efficiently locating objects using Hausdorff distance[J].International Journal of Computer Vision, 1997, 24(3):251-270.
[41]Qiu,Z.M.,Li,J.,Ge,J.,Zhou.Q.B.:The research of automatic object recognizing algorithms based on Hausdorff distance[J]. Infrared technology, 2006, 8(4):199-202.
[42]Comaniciu,D.,Meer P. R.:Analysis of Feature Spaces: Color image segmentation[C].In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition CVPR’97, San Juan, Puerto Rico, 1997:750-755.
[43]Zhang,Y.H.,Wei,W.,Yu,D.: Kalman tracking algorithm based on real-time vision of ping-pong robot[J].Zhejiang Daxue Xuebao, 2009 ,43( 9): 1580-1584
[44]Allen,J.G.,Richard,Y. D. J.:Object tracking using Camshaft algorithm and multiple quantized featurespaces[C].In: Pan-Sydney Area Workshop on Visual Information Processing VIP, 2003, Sydney, Australia, 2003:1-5.
[45]Peng,J.,Gu,L.Z.,Su,J.B.:Hand tracking for humanoid robot using Camshift algorithm and Kalman filter[J]. Journal of Shanghai Jiaotong University, 2006, 40(7):1161-1165.
[46]Poritza,B.:Hidden Markov models: a guided tour[C].In:Proc. of IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE, 1988: 7213.
[47]Vinod,N.,James,J.:Automated visual surveillance using Hidden Markov Models[C].In:the 15th Vision Interface Conference, Calgary, 2002:88-92.
[48]Chen,T., Huang,C.,Chang,C.Wang,J.:On the use of Gaussian mixture model for speaker variability analysis[C].In:Int. Conf. SLP, Denver, CO, 2002: 1249-1252
[49]De La Torre, F.,Black, M.J.: Robust principal component analysis for computer vision[C].In: Proceedings of the IEEE International Conference on Computer Vision, 2001:(1), 362-369.
[50]何强,何英.MATLAB扩展编程[M].清华大学出版社,2002:349-380.
[51]Rabiner,L.R.:Tutorial on hidden Markov models and selected applications in speech recognition[C].In: Proceedings of the IEEE, 1989, 77(2):257-286.
[52]Bashir,F.,Khokhar,A.,Schonfeld,D.:Automatic objects trajectory-based motion recognition using Gaussian mixture models[C].In: Proc of IEEE International Conference on Multimedia & Expt, Amsterdam,2005.
[53]国立新,莫福源,李昌立.基于连续高斯混合密度HMM的汉语全音节语音识研究[J].声学学报,1995.
[54]Lou,H.L.:Implementing the Viterbi algorithm[J].IEEE Signal Processing Magazine,1995, 12(5):42-52.
[55]Zhang,X.,Mason,J.S.:Improved training using semi-hidden Markov models in speech recognition[C]. In:ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing, 1989, 1:306-309.
[56]Kevin,M.:Hidden markov model toolbox for matlab.1998. [http://www.cs. ubc.ca/~murphyk/ Software/HMM/hmm.html ].
[57]Hao,J.Y.,Hao,S., Li,C.,Xiong,Z.,et al:Vehicle behavior understanding based on movement string[C]. In:IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC, 2009:243-248.
[58] Jolly,M. P. D., Jain ,A. K.:A modified hausdorff distance for object matching[C].In: In International Conference on Pattern Recognition, 1994, 6:566-568.
[2]王亮,胡卫明,谭铁牛.人运动的视觉分析综述[J].计算机学报, 2002,25(3):225-237.
[3]纪运红.智能监控系统若干关键技术的研究和实现[D].辽宁科技大学.2007.
[4]Buzan,D.,Sclaroff,S.,Kollios,G.:Extraction and clustering of motion trajectories in video[A]. In: Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, UK, 2004:521-524.
[5]Lee,K.K.,Yu,M.,Xu,Y.S.:Modeling of human walking trajectories for surveillance[C]. In: Proceedings of IEEE International Conference on Intelligent Robots and Systems, 2003, 2:1554-1559.
[6]Makris,D.,Ellis,T.:Learning semantic scene models from observing activity in visual surveill -ance[C].In:IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 2005, 35(3):397-408.
[7]胡卫明,谢丹,谭铁牛.轨迹分布模式学习的层次自组织神经网络方法[J].计算机学报,2003, 26(4):417-426.
[8]Johnson,N.,Hogg,D.:Learning the distribution of objects trajectories for event recognition[J].Image and Vision Computing, 1996, 14(8):609-615.
[9] Owens,J., Hunter,A.:Application of the self-organizing map to trajectory classification[C].In:Third IEEE International Workshop on Visual Surveillance, 2000, p77-83.
[10]Piciarelli,C.,Foresti,G.L.:Online trajectory clustering for anomalous events detection[J].Pattern Recog -nition Letters, 2006, 27(15):1835-1842.
[11]Li,X.,Hu,W.M.,Hu,W.:A coarse-to-fine strategy for vehicle motion trajectory clustering[A].In:Proceedings International Conference on Pattern Recognition, 2006, 1:591-594.
[12]Hu,W.,Xiao,X.,Xie,D.,Tan,T.and Maybank,S.:A system for learning statistical motion patterns[J]. IEEE Trans. Pattern Anal. Machine Intell, 2006, 28(9):1450-1464.
[13]Naftel,A. Khalid,S.:Classifying spatiotemporal object trajectories using unsupervised learning in the coefficient feature space[J]. ACM Multimedia Systems, 2006, 12(3):227-238.
[14]Yu,J.W.,Li,S.,Chen,Y.S:Trajectory-based video retrieval by string matching [C].In:Proceedings International Conference on Image Processing, ICIP, 2004,4:2243-2246.
[15]Stauffer,C.,Grimson,W.,Eric,L.:Learning patterns of activity using real-time tracking[J].Transactions onPattern Analysis and Machine Intelligence, IEEE, 2000, 22(8):747-757.
[16]Hu,W.M.,Xiao,X.J.,Xie,D.,Tan,T.N.,Maybank,Steve.:Traffic accident prediction using 3-D model-based vehicle tracking[J].Transactions on Vehicular Technology, IEEE, 2004, 53(3):677-694.
[17]Naftel,A.,Khalid,S.:Classifying spatiotemporal object trajectories using unsupervised learning in the coefficient feature space[J].Multimedia Systems, 2006, 12(3):227-238.
[18]Brand,M.,Oliver,N.,Pentland,A.:Coupled hidden Markov models for complex action recognition [C].In:Proc. of IEEE CS Conf. on Computer Vision and Pattern Recognition, 1997:994-999.
[19]Zhou,H.,Don,K.:Unusual event detection via multi-camera video mining[C].In:Proc. of the 18th International Conf. on Pattern Recognition, ICPR archive,2006, 3:1161-1166.
[20]李和平,胡占义,胡毅红,吴福朝.基于半监督学习的行为建模与异常检测[J].Jornal of Software,2007, 18(3):527-537.
[21]Vinod,N.,James,J.C.:Automated visual surverllance using hidden marknov models[J].Centre for Intelligent Machines, Mcgill University, Montreal, PQH3A2A7.2002.
[22]Naohiko,S.,Kosuke,H.,Kenichi,T.,Yoshinori,K.,Yoichi,S.:Learning motion patterns and anomaly detection by human trajectory analysis[C]. In: IEEE International Conference on Digital Object Identifie.2007: 498-503
[23]Morris,B.T.,Trivedi,M.M.:A survey of vision-based trajectory learning and analysis for surveillance [J].IEEE Transactions on Circuits and Systems for Video Technology, 2008, 18(8):1114-1127.
[24]Bashir,F.,Khokhara,S.D.:Automatic object trajectory-based motion recognition using gaussian mixture models[C].In:International Conference on Multimedia&Expt, IEEE 2005:1532-1535
[25]潘奇明,程咏梅.基于隐马尔科夫模型的运动目标轨迹识别[J].计算机应用研究,2008,25(7)
[26]张丽君,吴晓娟,盛赞.基于HMM复杂场景下的行为识别方法[J].2008,(7):07-076.
[27]于殿泓.图像检测与处理技术[M].西安电子科技大学出版社.2007.1:100-175.
[28]Verri,A.,Uars,S., DeMieheli,E.:Motion segmentation from optical flow[C].In: Proc. the 5th Alvey Vision Conf., Brighton, UK, 1989:209-214.
[29]Barron,J.,Fleet,D.,Beauehemin,S.:Performance of optical flow techniques[J].International Journal of Computer Vision, 1994:12(l):42-77.
[30]Tong,K.,Yun,J.H., et al .:Motion estimation using new blocks based on the frame difference for frame rate-up conversion[C].In:Proceedings of International Meeting on Information Display, 2008, 8:1043-1046.
[31]Zhan,C.H.,Duan,X.H.,Xu,S.Y.,et al.:An improved moving object detection algorithm based on frame difference and edge detection[C].In:Proceedings of the 4th International Conference on Image and Graphics, ICIG, 2007:519-523.
[32]Mahadevan,V.,Vasconcelos,N.:Background subtraction in highly dynamic scenes[C].In:26th IEEE Conference on Computer Vision and Pattern Recognition CVPR, 2008:1-6
[33]Thongkamwitoon,T.,Aramvith,S.,Chalidabhongse,T.H. :An adaptive real-time background subtraction and moving shadows detection[C].In:IEEE International Conference on Multimedia and Expo (ICME), 2004, 2:1459-1462.
[34]马义德,朱望飞,安世霞,邱会银,汤书森.改进的基于高斯混合模型的运动目标检测方法[J].计算机应用, 2007,27(10):2544-2546.
[35]Zivkovic,Z.:Improved adaptive Gaussian mixture model for background subtraction[A].In:Proceedings International Conference on Pattern Recognition, 2004, 2:28-31.
[36]Cheng,J.,Yang,J.,Zhou,Y.:A novel adaptive Gaussian mixture model for background subtraction[C].Pattern Recognition and Image Analysis: Second Iberian Conference, Lecture Notes in Computer Science, 2005, 3522(I):587-593.
[37]郝彦彬.基于视频序列图像的运动目标检测与跟踪[D].大连理工大学.2008.
[38]David,B.,Philip,M.,et al.:A Real-time Computer Vision System for Measuring Traffic Parameters[J]. Computer Vision and Pattern Recognition, IEEE, 1997:495-501
[39]Huang,F.Z.,Su,J.B.:Abstracting and tracking of outline of face based on level set method[J].Computer transaction, 2003, 26(4):491-496
[40]Rucklidge W. J.:Efficiently locating objects using Hausdorff distance[J].International Journal of Computer Vision, 1997, 24(3):251-270.
[41]Qiu,Z.M.,Li,J.,Ge,J.,Zhou.Q.B.:The research of automatic object recognizing algorithms based on Hausdorff distance[J]. Infrared technology, 2006, 8(4):199-202.
[42]Comaniciu,D.,Meer P. R.:Analysis of Feature Spaces: Color image segmentation[C].In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition CVPR’97, San Juan, Puerto Rico, 1997:750-755.
[43]Zhang,Y.H.,Wei,W.,Yu,D.: Kalman tracking algorithm based on real-time vision of ping-pong robot[J].Zhejiang Daxue Xuebao, 2009 ,43( 9): 1580-1584
[44]Allen,J.G.,Richard,Y. D. J.:Object tracking using Camshaft algorithm and multiple quantized featurespaces[C].In: Pan-Sydney Area Workshop on Visual Information Processing VIP, 2003, Sydney, Australia, 2003:1-5.
[45]Peng,J.,Gu,L.Z.,Su,J.B.:Hand tracking for humanoid robot using Camshift algorithm and Kalman filter[J]. Journal of Shanghai Jiaotong University, 2006, 40(7):1161-1165.
[46]Poritza,B.:Hidden Markov models: a guided tour[C].In:Proc. of IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE, 1988: 7213.
[47]Vinod,N.,James,J.:Automated visual surveillance using Hidden Markov Models[C].In:the 15th Vision Interface Conference, Calgary, 2002:88-92.
[48]Chen,T., Huang,C.,Chang,C.Wang,J.:On the use of Gaussian mixture model for speaker variability analysis[C].In:Int. Conf. SLP, Denver, CO, 2002: 1249-1252
[49]De La Torre, F.,Black, M.J.: Robust principal component analysis for computer vision[C].In: Proceedings of the IEEE International Conference on Computer Vision, 2001:(1), 362-369.
[50]何强,何英.MATLAB扩展编程[M].清华大学出版社,2002:349-380.
[51]Rabiner,L.R.:Tutorial on hidden Markov models and selected applications in speech recognition[C].In: Proceedings of the IEEE, 1989, 77(2):257-286.
[52]Bashir,F.,Khokhar,A.,Schonfeld,D.:Automatic objects trajectory-based motion recognition using Gaussian mixture models[C].In: Proc of IEEE International Conference on Multimedia & Expt, Amsterdam,2005.
[53]国立新,莫福源,李昌立.基于连续高斯混合密度HMM的汉语全音节语音识研究[J].声学学报,1995.
[54]Lou,H.L.:Implementing the Viterbi algorithm[J].IEEE Signal Processing Magazine,1995, 12(5):42-52.
[55]Zhang,X.,Mason,J.S.:Improved training using semi-hidden Markov models in speech recognition[C]. In:ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing, 1989, 1:306-309.
[56]Kevin,M.:Hidden markov model toolbox for matlab.1998. [http://www.cs. ubc.ca/~murphyk/ Software/HMM/hmm.html ].
[57]Hao,J.Y.,Hao,S., Li,C.,Xiong,Z.,et al:Vehicle behavior understanding based on movement string[C]. In:IEEE Conference on Intelligent Transportation Systems, Proceedings, ITSC, 2009:243-248.
[58] Jolly,M. P. D., Jain ,A. K.:A modified hausdorff distance for object matching[C].In: In International Conference on Pattern Recognition, 1994, 6:566-568.