Detection and recognition of sports(wo)men from multiple views

The methods presented in this paper aim at detecting and recognizing players on a sport-field, based on a distributed set of loosely synchronized cameras. Detection assumes player verticality, and sums the cumulative projection of the multiple views' foreground activity masks on a set of planes that are parallel to the ground plane. After summation, large projection values indicate the position of the player on the ground plane. This position is used as an anchor for the player bounding box projected in each one of the views. Within this bounding box, the regions provided by mean-shift segmentation are sorted out based on contextual features, e.g. relative size and position, to select the ones that are likely to correspond to a digit. Normalization and classification of the selected regions then provides the number and identity of the player. Since the player number can only be read when it faces towards the camera, graph-based tracking is considered to propagate the identity of a player along its trajectory.