Put Me In, Coach!
Automating video storytelling for sports.
Sports visualization has come a long way with the growth of commercial elements of sports attached to it. In the last two decades, we have seen sports visualization go from just showcasing data tables on the broadcast to showing in-depth augmented visualizations that span from slow-motion shots to spotlights and maps on the playing field.
The biggest challenge with creating sports visualization for broadcast or analysis is that you need considerable expertise in video editing. A recent paper by Chen et al. addresses this challenge. The authors have successfully created a tool for visualizing visualize the data for a sport such as table tennis using a sequence of simple mouse clicks.
Let’s say you want to analyze a particular segment of a table tennis match to be shown to the players and coaches for understanding some nuances of the game. You have footage from a particular match that is ready to be analyzed. Load it up into the VisCommentator and get ready to click to augment your videos using four simple steps.
Brush the timeline: The analyst can use the timeline as in a conventional video editor to find specific highlights in the match, such as where the ball rotation speed is so fast that the opponent can only return it to a narrow area.
Select the Data: The tool automatically tags the video and directly shows the events to visualize. To look into player strokes, the analyst right-clicks to get automatic data suggestions (ball position, placement, rotation speed, potential placement, and potential routes).
Select the purpose and order: The analyst can now choose clips and order them appropriately. There are two basic purposes of sports data: educational (low semantics) and entertainment (high semantics). As for order, not all sports video analysis is chronological. The authors talk about six ordering methods: linear, flashforward, zigzag, flashback, timefork, and grouped. Based on these the tool automatically suggests a possible mapping between purpose and order for the analyst to select and visualize.
Edit and Fine-tune: Finally the analyst can cut and edit the exact parts required to build a final product that can be showcased to the coaching staff as well as the players.
What I found interesting about this tool was that it actively helps analysts to smoothen the storytelling process. They can build a storyline without wondering about the technicalities of editing a video and showcasing different visual channels on it. This way, storytelling can be made accessible even to people with no specialized knowledge in these skills.
References
- Z. Chen et al., “Augmenting Sports Videos with VisCommentator,” in IEEE Transactions on Visualization and Computer Graphics, vol. 28, no. 1, pp. 824–834, Jan. 2022, doi: 10.1109/TVCG.2021.3114806.
