Improved Qualitative Trajectory Calculus for Pair-Activity Analysis
|
摘要
Trajectory provides the most robust feature for activity recognition in far-field surveillance videos,in which increasing attentions have been given to the use of qualitative methods with symbolic rather than real-value features. Qualitative trajectory calculus( QTC) showed a good performance in pair-activity from video. However,QTC and similar works are not good at dealing with noise,since they are all considering short-term features. To deal with the problems mentioned above,two types of long-term features,including sub-trajectory feature and point-trajectory feature,are designed. The sub-trajectory feature is a long-term feature in a coarse granularity,while the point-trajectory feature is a long-term feature in a relatively fine granularity. Using the sub-trajectory feature,a couple of trajectories are segmented into sub-trajectories and enveloping boxes are used to substitute the original sub-trajectory for capturing the major attributes. The point-trajectory feature describes the relationship between a single point in one trajectory and all parts of the other trajectory. The experiments on the human activity classification data demonstrated that our proposed methods are better than the original QTC and previous short-term features.
Trajectory provides the most robust feature for activity recognition in far-field surveillance videos,in which increasing attentions have been given to the use of qualitative methods with symbolic rather than real-value features. Qualitative trajectory calculus( QTC) showed a good performance in pair-activity from video. However,QTC and similar works are not good at dealing with noise,since they are all considering short-term features. To deal with the problems mentioned above,two types of long-term features,including sub-trajectory feature and point-trajectory feature,are designed. The sub-trajectory feature is a long-term feature in a coarse granularity,while the point-trajectory feature is a long-term feature in a relatively fine granularity. Using the sub-trajectory feature,a couple of trajectories are segmented into sub-trajectories and enveloping boxes are used to substitute the original sub-trajectory for capturing the major attributes. The point-trajectory feature describes the relationship between a single point in one trajectory and all parts of the other trajectory. The experiments on the human activity classification data demonstrated that our proposed methods are better than the original QTC and previous short-term features.
Trajectory provides the most robust feature for activity recognition in far-field surveillance videos,in which increasing attentions have been given to the use of qualitative methods with symbolic rather than real-value features. Qualitative trajectory calculus( QTC) showed a good performance in pair-activity from video. However,QTC and similar works are not good at dealing with noise,since they are all considering short-term features. To deal with the problems mentioned above,two types of long-term features,including sub-trajectory feature and point-trajectory feature,are designed. The sub-trajectory feature is a long-term feature in a coarse granularity,while the point-trajectory feature is a long-term feature in a relatively fine granularity. Using the sub-trajectory feature,a couple of trajectories are segmented into sub-trajectories and enveloping boxes are used to substitute the original sub-trajectory for capturing the major attributes. The point-trajectory feature describes the relationship between a single point in one trajectory and all parts of the other trajectory. The experiments on the human activity classification data demonstrated that our proposed methods are better than the original QTC and previous short-term features.
引文
[1]Zhang W,Chen F,Xu W,et al.Hierarchical group process representation in multi-agent activity recognition[J].Signal Processing:Image Communication,2008,23(10):739-753.
[2]Liu C,Chung Y,Chung P,et al.An interaction-embedded HMM framework for human behavior understanding:with nursing environments as examples[J].IEEE Transactions on Information Technology in Biomedicine,2010,14(5):1236-1246.DOI:10.1109/TITB.2010.2052061.
[3]Liu X,Chua C S.Multi-agent activity recognition using observation decomposed hidden Markov model[C]∥International Conference on Computer Vision Systems.Heidelberg,Berlin:Springer,2003:247-256.
[4]Guo P,Miao Z.Multi-person activity recognition through hierarchical and observation decomposed HMM[C]∥Multimedia and Expo(ICME),2010 IEEE International Conference on,IEEE,2010:143-148.
[5]Dore A,Regazzoni C S.Interaction analysis with a Bayesian trajectory model[J].IEEE Intelligent Systems,2010,25(3):32-40.
[6]Sato K,Aggarwal J K.Tracking and recognizing twoperson interactions in outdoor image sequences[J].Current Opinion in Organ Transplantation,2001:87-94.
[7]Hu T,Zhu X,Guo W,et al.Efficient interaction recognition through positive action representation[J].Mathematical Problems in Engineering,2013:1-11.
[8]Messing R,Pal C,Kautz H.Activity recognition using the velocity histories of tracked keypoints[C]∥Computer Vision,2009 IEEE 12th International Conference on,IEEE,2009:104-111.
[9]Sun J,Wu X,Yan S,et al.Hierarchical spatio-temporal context modeling for action recognition[C]∥Computer Vision and Pattern Recognition,CVPR 2009,IEEE Conference on,IEEE,2009:2004-2011.
[10]Wang X,Ma K T,Ng G W,et al.Trajectory analysis and semantic region modeling using nonparametric hierarchical Bayesian models[J].International Journal of Computer Vision,2011,95(3):287-312.
[11]AlZoubi A,Al-Diri B,Pike T,et al.Pair-activity analysis from video using qualitative trajectory calculus[J].IEEE Transactions on Circuits and Systems for Video Technology,2018,28(8):1850-1863.
[12]De Weghe N V,Cohn A G,De Tre G,et al.A qualitative trajectory calculus as a basis for representing moving objects in geographical information systems[J].Control and Cybernetics,2006,35(1):97-119.
[13]Nico V D W,Cohn A G,Maeyer P D,et al.Representing moving objects in computer-based expert systems:the overtake event example[J].Expert Systems with Applications,2005,29(4):977-983.
[14]Hanheide M,Peters A,Bellotto N.Analysis of humanrobot spatial behaviour applying a qualitative trajectory calculus[C]∥RO-MAN,2012 IEEE,2012:689-694.
[15]Sethi R J,Jo H,Gil Y.Structured analysis of the ISIatomic pair actions dataset using workflows[J].Pattern Recognition Letters,2013,34(15):2023-2032.(Wang)
[2]Liu C,Chung Y,Chung P,et al.An interaction-embedded HMM framework for human behavior understanding:with nursing environments as examples[J].IEEE Transactions on Information Technology in Biomedicine,2010,14(5):1236-1246.DOI:10.1109/TITB.2010.2052061.
[3]Liu X,Chua C S.Multi-agent activity recognition using observation decomposed hidden Markov model[C]∥International Conference on Computer Vision Systems.Heidelberg,Berlin:Springer,2003:247-256.
[4]Guo P,Miao Z.Multi-person activity recognition through hierarchical and observation decomposed HMM[C]∥Multimedia and Expo(ICME),2010 IEEE International Conference on,IEEE,2010:143-148.
[5]Dore A,Regazzoni C S.Interaction analysis with a Bayesian trajectory model[J].IEEE Intelligent Systems,2010,25(3):32-40.
[6]Sato K,Aggarwal J K.Tracking and recognizing twoperson interactions in outdoor image sequences[J].Current Opinion in Organ Transplantation,2001:87-94.
[7]Hu T,Zhu X,Guo W,et al.Efficient interaction recognition through positive action representation[J].Mathematical Problems in Engineering,2013:1-11.
[8]Messing R,Pal C,Kautz H.Activity recognition using the velocity histories of tracked keypoints[C]∥Computer Vision,2009 IEEE 12th International Conference on,IEEE,2009:104-111.
[9]Sun J,Wu X,Yan S,et al.Hierarchical spatio-temporal context modeling for action recognition[C]∥Computer Vision and Pattern Recognition,CVPR 2009,IEEE Conference on,IEEE,2009:2004-2011.
[10]Wang X,Ma K T,Ng G W,et al.Trajectory analysis and semantic region modeling using nonparametric hierarchical Bayesian models[J].International Journal of Computer Vision,2011,95(3):287-312.
[11]AlZoubi A,Al-Diri B,Pike T,et al.Pair-activity analysis from video using qualitative trajectory calculus[J].IEEE Transactions on Circuits and Systems for Video Technology,2018,28(8):1850-1863.
[12]De Weghe N V,Cohn A G,De Tre G,et al.A qualitative trajectory calculus as a basis for representing moving objects in geographical information systems[J].Control and Cybernetics,2006,35(1):97-119.
[13]Nico V D W,Cohn A G,Maeyer P D,et al.Representing moving objects in computer-based expert systems:the overtake event example[J].Expert Systems with Applications,2005,29(4):977-983.
[14]Hanheide M,Peters A,Bellotto N.Analysis of humanrobot spatial behaviour applying a qualitative trajectory calculus[C]∥RO-MAN,2012 IEEE,2012:689-694.
[15]Sethi R J,Jo H,Gil Y.Structured analysis of the ISIatomic pair actions dataset using workflows[J].Pattern Recognition Letters,2013,34(15):2023-2032.(Wang)