We solicit submissions for the special session on "Analysis of Group Interactions", held as part of the FG2017 main conference. We intend to accept five papers covering the broad scope of the topic.
Motivation
It is in our interaction with others that we give meaning to our individual as well as collective behavior. Dyadic (two-person) and small group interactions punctuate our everyday life, when we participate in work meetings, when we encounter acquaintances on the street, when we move in emergent groups, and so on. Group interaction amounts to far more than the collection of individual behaviors. Coordination in both time and space is often meaningful for the specific interaction setting or the outcome of the interaction.
In the past decades, we witnessed impressive advances in the visual analysis of face and body. We learned how to robustly and accurately track facial expressions and body movements, and thus started to interpret individual behavior. With the solid base from this research, it's time to take the next step and move beyond the analysis of the individual. In the special session "Analysis of Group Interactions" at FG2017, we focus on challenges in three areas.
First, we consider technological challenges in the detection and recognition of dyadic and group interactions based on the analysis of face and body. Often, in fact, interactions are observed under unfavorable conditions, with occlusions and challenging viewpoints. At the same time, the presence of others provides cues that can be exploited to achieve more robust results. The detection and recognition of behavior coordination is another challenge.
Second, we focus on the investigation of affect and personality starting from the behavior of members of a group. An example is the analysis of collective mood from the behaviors of group members. We also face challenges in the investigation how personality and affect shape interactions.
Third, we focus on the analysis of group concerted actions, and face challenges in the analysis of group dynamics. The focus on temporal patterns of behavior will bring novel insights and applications, but requires us to effectively model the state of the interaction. We also consider the analysis of the (dis)assembly of groups, and the identification of groups within crowds based on observed behavior.
While these challenges all contain a technological component, we advocate a broad, multi-disciplinary perspective on group analysis. We acknowledge the technical advances in computer science to analyze behavior from video and other sensors. We explicitly also foresee an important role for the social sciences. We believe that work that will stand the test of time will combine engineering and theoretical knowledge effectively.
Submission
We solicit high-quality theoretical, empirical and application research papers on the topics of the special session, ideally with a multi-disciplinary perspective. Papers, according to the FG guidelines, can be submitted through the Content Management System, and then select the "Group Analysis" track. After peer review, accepted papers will be published in the main conference proceedings on IEEE Xplore.
Topics
- Detecting and recognizing interactions
- Computer vision
- Novel cues for interaction analysis
- Interaction forecasting
- Social role recognition
- Models for turn-taking
- Affect and personality in group interaction
- Recognition of affect from interactions
- Influence of personality on interaction
- Group behavior and dynamics
- Abnormal group behavior recognition
- Assembly and disassembly of groups
- Group behavior analysis at different ages
- Group outcome prediction
Deadlines
- Paper submission: extended to December 22, 2016
- Notification of acceptance: January 23, 2017
- Camera-ready papers: Februari 17, 2017
- FG conference: May 31 - June 2, 2017
Organizers
- Ronald Poppe - Utrecht University
- Marco Cristani - University of Verona
- Chiara Bassetti - University of Trento