An Effective Approach Based on a Subset of Skeleton Joints for Two-Person Interaction Recognition

In this article, the problem of human interaction recognition is addressed. A novel methodology is presented to model the interaction between two people using a Kinect sensor. It is proposed to analyse the distances between a subset of skeleton joints to determine their contribution to the recognition of interaction. Subsequently, the most representative joints are taken into account for the formation of a pentagon for each person. Five Euclidean distances are extracted between the vertices of two pentagons corresponding to the people to analyse. Finally, the SVM is used for the recognition of human interaction. Experimental results demonstrate the effectiveness of this work compared to recently published proposals.

[1]  James M. Rehg,et al.  Decoding Children's Social Behavior , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  Wanqing Li,et al.  Action recognition based on a bag of 3D points , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Workshops.

[3]  Guijin Wang,et al.  A novel hierarchical framework for human action recognition , 2016, Pattern Recognit..

[4]  Mohan M. Trivedi,et al.  Multi-person interaction and activity analysis: a synergistic track- and body-level analysis framework , 2007, Machine Vision and Applications.

[5]  Changyin Sun,et al.  Action Recognition Using Nonnegative Action Component Representation and Sparse Basis Selection , 2014, IEEE Transactions on Image Processing.

[6]  Adolfo López,et al.  Connected operators on 3D data for human body analysis , 2011, CVPR 2011 WORKSHOPS.

[7]  Mahmoud Neji,et al.  A Framework for People Re-Identification in Multi-Camera Surveillance Systems. , 2017 .

[8]  Hong Cheng,et al.  Learning contrastive feature distribution model for interaction recognition , 2015, J. Vis. Commun. Image Represent..

[9]  Dimitris Samaras,et al.  Two-person interaction detection using body-pose features and multiple instance learning , 2012, 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops.

[10]  Honghai Liu,et al.  Human-human interaction recognition based on spatial and motion trend feature , 2017, 2017 IEEE International Conference on Image Processing (ICIP).

[11]  Bingbing Ni,et al.  RGBD-HuDaAct: A color-depth video database for human daily activity recognition , 2011, ICCV Workshops.

[12]  Hong Cheng,et al.  Interactive body part contrast mining for human interaction recognition , 2014, 2014 IEEE International Conference on Multimedia and Expo Workshops (ICMEW).

[13]  Leonid Sigal,et al.  Poselet Key-Framing: A Model for Human Activity Recognition , 2013, 2013 IEEE Conference on Computer Vision and Pattern Recognition.

[14]  Junxia Gu,et al.  Action and Gait Recognition From Recovered 3-D Human Joints , 2010, IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics).

[15]  Robert B. Fisher,et al.  The BEHAVE video dataset: ground truthed video for multi-person behavior classification , 2010 .

[16]  Yunde Jia,et al.  Learning Human Interaction by Interactive Phrases , 2012, ECCV.

[17]  Hongqi Wang,et al.  Human Interaction Recognition Based on Transformation of Spatial Semantics , 2012, IEEE Signal Processing Letters.

[18]  Ankur Agarwal,et al.  Recovering 3D human pose from monocular images , 2006, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[19]  Hongqi Wang,et al.  Image Classification Based on pLSA Fusing Spatial Relationships Between Topics , 2012, IEEE Signal Processing Letters.

[20]  Jake K. Aggarwal,et al.  Semantic Representation and Recognition of Continued and Recursive Human Activities , 2009, International Journal of Computer Vision.