By Kenji Okuma
The work addresses the problem of automatically analyzing hockey scenes by estimating the panning, tilting, and zooming parameters of the broadcasting cameras, tracking hockey players in these scenes, and constructing a visual description of the scenes as trajectories of those players. Given quite fast and non-smooth camera motions to capture highly complex and dynamic scenes of hockey, tracking hockey players that are small blob-like, non-rigid and amorphous becomes an extremely difficult task. Although the domain of the problem in this work is hockey, our approach is applicable to a wider range of problems such as surveillance or any situations that require object tracking on a planar surface. The first part of the talk is about how to estimate homography (i.e., plane projective transformation) and extract camera motion. The second half of the talk explains how to track hockey players. The result demonstrates a quite successful performance of our system.
Back to the LCI Forum page