Jason Shang, Marjan Shahi

June 6, 2023, 11:30 am, EC4-2101A

*** 11am Presenter: Jason Shang ***

Hockey is a fast-paced sport, with a massive following, and games are often analyzed in order to determine various statistics and ways to improve players and teams. Much of this analysis requires location information of players on the ice, and with the absence of camera parameters, alternative methods are needed to obtain this. One method is homography warping for rink registration, where broadcast video frames are warped onto an overhead template, allowing for everything to be analyzed in a shared reference frame. In this seminar, we examine ways to improve this process as well as extend this method to work on alternate rink types using domain adaptation and synthetic data. 

*** 11:3am Presenter: Marjan Shahi ***

In the field of computer vision-driven ice hockey analytics, one of the most challenging and least studied tasks is goalie pose estimation. Unlike general human pose estimation, goalie pose estimation is much more complex as it involves not only the detection of keypoints corresponding to the joints of the goalie concealed under thick padding and mask, but also a large number of non-human keypoints corresponding to the large leg pads and gloves worn, the stick, as well as the hockey net. To tackle this challenge, we introduce GoalieNet, a multi-stage deep neural network for jointly estimating the pose of the goalie, their equipment, and the net. Experimental results using NHL benchmark data demonstrate that the proposed GoalieNet can achieve 70\% accuracy across all keypoints, which indicates that such a joint pose estimation approach can be a promising direction for tackling this challenge.