Characterization of Social Interactions

<< Back to Research Areas & Projects

We are developing methods that can detect and recognize social interactions. Structures formed in our daily interactions play very important roles in understanding the social behavior, such as in office meetings, social gatherings, festivals and memorable events. In this work, we employ uncalibrated first person (or egocentric) views, face and body features analysis to detect and classify structures formed in interactions. Our focus is on real-world videos, like meetings in office environment, a trip to am amusement park or in scenarios where large number of people are involved. We proposed a search-based structure recovering method to reconstruct the complete social interaction with multiple .first-person views, where these views contributed to the multifaceted understanding of the social interaction (as shown in the figure below).

It populates the interaction group with multiple views, which are useful to present a complete interaction structure. We proposed a search-based reconstruction method, which are simpler than 3D reconstruction yet useful in capturing the social interaction spatial structure. To the best of our knowledge, this is the first time the multiple first-person view cameras are combined to analyze the social interaction spatial structure.

Details read here:

[C4] G. Tian, Y. Wong, B. Mandal, V. Chandrasekha and M. Kankanhalli, "Multi-sensor Self-Quantification of Presentations", ACM Multimedia (ACMMM), Brisbane, Australia, Oct 2015. (Accepted Oral) [PDF]

[C3] S. Ching, B. Mandal, Q. Xu, Li Liyuan and J-H. Lim, "Enhancing Social Interaction with Seamless Face Recognition on Google Glass: Leveraging opportunistic multi-tasking on smart phones", 17th International Conference on Human-Computer Interaction with Mobile Devices and Services (MobileCHI) August 24th - 27th, 2015 Copenhagen, Denmark. (Accepted) [PDF]

[C2] Q. Xu, M. Mukawa, Liyuan Li, J-H Lim, C. Tan, S. Ching, T. Gan and B. Mandal, "Exploring Users Attitudes towards Social Interaction Assistance on Google Glass", ACM 6th International Conference on Augmented Human (AH), pp. 9-12, Mar 2015. [PDF]

[C1] G. Tian, Y. Wong, B. Mandal, V. Chandrasekhar, Liyuan Li, J-H Lim and M. Kankanhalli, "Recovering Social Interaction Spatial Structure from Multiple First-person Views", 3rd International Workshop on Socially-Aware Multimedia, ACM Multimedia, Orlando (FL), USA, pp. 7-12, Nov 2014. [PDF]

<< Back to Research Areas & Projects