Figure 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Subset of 1500 puck locations in the dataset. The puck locations on the ice rink are highly correlated with the event label. Faceoffs(red) are located at the faceoff circles, shots(blue) are located in the offensive zones and dump in/outs (yellow) are presents in the neutral zone.
Source publication
Puck localization is an important problem in ice hockey video analytics useful for analyzing the game, determining play location, and assessing puck possession. The problem is challenging due to the small size of the puck, excessive motion blur due to high puck velocity and occlusions due to players and boards. In this paper, we introduce and imple...
Contexts in source publication
Context 1
... on the test set. The network is able to localize the puck during player and board occlusions. At test-time, the network is able to perform inference using a sliding window approach in previously unseen untrimmed broadcast hockey video at 5 frame per second (fps). Player and puck location information is related with event occurring on the rink (Fig. 1). Other research leverages player and ball trajectories for event recognition using a separate tracking/localization system [11,17]. We attach an event recognition head to the puck localization model to leverage the puck location information for event recognition and train the whole network using multi-task learning. Experimental ...
Context 2
... videos are split into 80% samples for training and 10% samples each for validation and testing. Fig 1 shows the distribution of a subset of puck location data. The videos are also annotated with an event label which can be either Faceoff, Advance (dump in/out), Play ( player moving the puck with an intended recipient e.g., pass, stickhandle ) or Shot. ...
Context 3
... lower stride results in noisy results and higher strides produces very simple predictions. The network is tested on another 10 second video with l = 2s and s = 1s shown in Fig 10. The predicted puck trajectory is shown in Fig 10. ...
Similar publications
Motion planning is a mature area of research in robotics with many well-established methods based on optimization or sampling the state space, suitable for solving kinematic motion planning. However, when dynamic motions under constraints are needed and computation time is limited, fast kinodynamic planning on the constraint manifold is indispensab...
El hockey es un deporte de conjunto que se encuentra en constante crecimiento, tanto en cantidad de jugadores/as, debido a la cosecha de grandes triunfos de sus primeros equipos, mayor difusión por parte de los medios de comunicación, inclusión del deporte en los programas de Educación Física; como así también de la calidad de sus jugadores. Tantos...
This study compared vitamin D sufficiency between indoor and outdoor elite athletes. We also evaluated the association between vitamin D level, body composition, and stress fractures incidence. 27 outdoor elite male collegiate athletes (field hockey players) and 21 indoor elite male collegiate athletes (fencing players) were enrolled. Participants'...
The paper aims at the evaluation of efficiency in sports. Many articles are dealing with the application of data envelopment analysis (DEA) models in this area. They are mainly oriented on efficiency evaluation of teams and not the individual players. On the contrary, the main aim of this paper is to combine both approaches and investigate the rela...
