Technical staff increasingly consider the player’s orientation as a determining characteristic in obtaining a positional advantage during a match. Optimal orientation to correspond with specific situations can be taught and learned during training.

Recent studies in team sports have focused on tactical analysis using player positioning, obtained both by optical tracking and by GPS; but, as far as we know, data regarding a player’s orientation have not yet been provided or analyzed from a quantitative point of view.

We have designed and implemented a sequential function for the field of sports analysis and technology, represented in Figure 1, which predicts the orientation of all players on the playing field through the use of a panoramic video.

Our function first splits the panoramic video into frames. Then, for each frame, it cuts out the players and for each cut-out it estimates the player’s 2D pose (skeleton) using an Artificial Intelligence technique, Machine Learning, which allows it to process information through the use of examples.

Once the 2D pose has been estimated, an orientation proposal is made for the trimming of that frame using another Artificial Intelligence technique. As a final step in our function, we combine a player’s orientations over a series of time intervals in order to create a more consistent orientation.

Incorporating the orientation of each player during the match would generate multiple benefits to improve current space-time analyses such as space control, pass probability, defensive pressure and other models that depend on players’ movement and positioning over time.

Our proposal has been evaluated both visually and numerically against a portable tracking system (RealTrack System), the data of which has already been validated.

Figure 2 shows the visual evaluation. The orientation of a random frame of six selected players (one center-back, one midfielder and one forward per team) is displayed. The green arrow corresponds to the orientation of the tracking device and the yellow arrow corresponds to the orientation predicted by our function.

In order to analyze numerically, we calculated, per minute, the number of meters that each player travels looking forward, backward and sideways. The graphs in Figure 3 show the differences between the two data sources.

The advantages of obtaining the orientation through the panoramic video instead of from the tracking device are that players do not have to carry any device on them, you can obtain the orientation of rival players and you can obtain the orientation of previously played matches where there was only video.

This type of analysis could help coaching staff identify player shortcomings in terms of body orientation relative to their position, or difficulties in the team’s ability to have more fluidity in the game.

Carlos Rodríguez

Sports Analysis and Technology, FC Barcelona