Y.: Tracking sports players with contextconditioned motion models

We employ hierarchical data association to track play-ers in team sports. Player movements are often complex and highly correlated with both nearby and distant players. A single model would require many degrees of freedom to represent the full motion diversity and could be difficult to use in practice. Instead, we introduce a set of Game Con-text Features extracted from noisy detections to describe the current state of the match, such as how the players are spa-tially distributed. Our assumption is that players react to the current situation in only a finite number of ways. As a result, we are able to select an appropriate simplified affin-ity model for each player and time instant using a random decision forest based on current track and game context fea-tures. Our context-conditioned motion models implicitly in-corporate complex inter-object correlations while remain-ing tractable. We demonstrate significant performance im-provements over existing multi-target tracking algorithms on basketball and field hockey sequences several minutes in duration and containing 10 and 20 players respectively. 1