Who knows who - Inverting the Social Force Model for finding groups

Abstract

Social groups based on friendship or family relations are very common phenomena in human crowds and a valuable cue for a crowd activity recognition system. In this paper we present an algorithm for automatic on-line inference of social groups from observed trajectories of individual people. The method is based on the Social Force Model (SFM)widely used in crowd simulation applications -which specifies several attractive and repulsive forces influencing each individual relative to the other pedestrians and their environment. The main contribution of the paper is an algorithm for inference of the social groups (parameters of the SFM) based on analysis of the observed trajectories through attractive or repulsive forces which could lead to such behaviour. The proposed SFM-based method shows its clear advantage especially in more crowded scenarios where other state-of-the-art methods fail. The applicability of the algorithm is illustrated on an abandoned bag scenario.

Online Bayesian Non-parametrics for Social Group Detection

“…Later, this model has been extended for dealing with groups [6]. Sochman and Hogg [25], on the other hand, propose a new agglomerative clustering method for group detection by inverting the SFM to infer its hidden parameters given tracking observations. Similarly, Ge et al [10] use agglomerative clustering to group together tracking trajectories gathered in a time-window with fixed length.…”

“…Among the works addressing the analysis of group behaviour, it is interesting to highlight the differences between 2 classes of methods, i.e., those considering only positional features at each time step [13,21] (e.g., the ground floor position), and those processing visual data over a time-window [3,10,25,28] (e.g., exploiting trajectories). The former provides the results at each time frame with no delay.…”

Weighted Interaction Force Estimation for Abnormality Detection in Crowd Scenes

“…The influence of moving particles in the same group should be zero [18]. We will use spectral clustering method to obtain the particle group information.…”

An Online Learned Elementary Grouping Model for Multi-target Tracking

“…Some work in computer vision [3][11] [24][28] has explored group discovery and group tracking, while our work focuses on using social groups to maintain individual identities in a DAT framework. T1 T2 T3 T5 T4 T6 T7 T8 T1 T2 T3 T5 T4 T6 T7 T8 Confident Tracklets T1-T3 T2-T3 T4-T6 T6-T7 T7-T8 g1 g2 g3 g4 g5 T2-T3 g2 T7-T8 g3 T1-T3 T2-T3 T4-T6 T6-T7 T7-T8 g1 g2 g3 g4 g5 Group Tracking…”