Group Activity Recognition using Belief Propagation for P2P Mobile Devices. (Technical Report, September 2nd 2013)

Human are social beings and spend most of their time in groups. Group behavior is emergent, generated by members’ personal characteristics and their interactions making it difficult to recognize in peer-to-peer (P2P) systems where the emergent behavior itself cannot be directly observed. We introduce 2 novel algorithms for distributed probabilistic inference (DPI) of group activities using loopy belief propagation (LBP). We evaluate their performance using an experiment in team sports activities and show that these activities are emergent in nature through natural processes. Centralized recognition performs very well, upwards of an F-score of 0.95 for large window sizes. The distributed methods iteratively converge to solutions which are comparable to centralized methods, even surpassing them in some situations. DPI-LBP also greatly reduces energy consumption of the node, where a centralized unit or infrastructure is not required, although memory and processor consumption increases.

[1]  Carlos Guestrin,et al.  Robust Probabilistic Inference in Distributed Systems , 2004, UAI.

[2]  Norman Dziengel,et al.  A system for distributed event detection in wireless sensor networks , 2010, IPSN '10.

[3]  Carlos Guestrin,et al.  A robust architecture for distributed inference in sensor networks , 2005, IPSN 2005. Fourth International Symposium on Information Processing in Sensor Networks, 2005..

[4]  Ling Bao,et al.  Activity Recognition from User-Annotated Acceleration Data , 2004, Pervasive.

[5]  Ting Yu,et al.  Group Level Activity Recognition in Crowded Environments across Multiple Cameras , 2010, 2010 7th IEEE International Conference on Advanced Video and Signal Based Surveillance.

[6]  K. D. Benne,et al.  Functional Roles of Group Members , 2010 .

[7]  Jian Lu,et al.  Mining Emerging Patterns for recognizing activities of multiple users in pervasive computing , 2009, 2009 6th Annual International Mobile and Ubiquitous Systems: Networking & Services, MobiQuitous.

[8]  Julia K. Parrish,et al.  Extracting Interactive Control Algorithms from Group Dynamics of Schooling Fish , 2005 .

[9]  Rama Chellappa,et al.  Learning multi-modal densities on Discriminative Temporal Interaction Manifold for group activity recognition , 2009, CVPR.

[10]  Jane Yung-jen Hsu,et al.  Probabilistic models for concurrent chatting activity recognition , 2009, TIST.

[11]  Paul Lukowicz,et al.  CoenoSense: A Framework for Real-Time Detection and Visualization of Collective Behaviors in Human Crowds by Tracking Mobile Devices , 2013 .

[12]  M. Smith.  Field Theory in Social Science: Selected Theoretical Papers. , 1951 .

[13]  Jian Lu,et al.  Recognizing multi-user activities using wearable sensors in a smart home , 2011, Pervasive Mob. Comput..

[14]  Michael Beigl,et al.  Global Peer-to-Peer Classification in Mobile Ad-Hoc Networks: A Requirements Analysis , 2011, CONTEXT.

[15]  Arun Venkataramani,et al.  Energy consumption in mobile phones: a measurement study and implications for network applications , 2009, IMC '09.

[16]  Kevin P. Murphy,et al.  Machine learning - a probabilistic perspective , 2012, Adaptive computation and machine learning series.

[17]  Michael Beigl,et al.  Towards Collaborative Group Activity Recognition Using Mobile Devices , 2013, Mob. Networks Appl..

[18]  Leah Edelstein-Keshet,et al.  Inferring individual rules from collective behavior , 2010, Proceedings of the National Academy of Sciences.

[19]  Dirk Helbing,et al.  Recognition of crowd behavior from mobile sensors with pattern analysis and graph clustering methods , 2011, Networks Heterog. Media.

[20]  Laurent Magnin,et al.  Elements about the Emergence Issue: A Survey of Emergence Definitions , 2006, Complexus.

[21]  Andreas Huth,et al.  THE SIMULATION OF FISH SCHOOLS IN COMPARISON WITH EXPERIMENTAL DATA , 1994 .

[22]  Alex Pentland,et al.  Using the influence model to recognize functional roles in meetings , 2007, ICMI '07.

[23]  D. Helbing,et al.  The Walking Behaviour of Pedestrian Social Groups and Its Impact on Crowd Dynamics , 2010, PloS one.