AISpotter uses computer vision to make coaches’ jobs easier

Watching video playback of a game’s key moments is an important part of coaching — but according to the team at AISpotter, it can also be tedious and time-consuming. That’s why they’re working to automate the process.

AISpotter took the stage today in Berlin as part of TechCrunch’s Startup Battlefield. CEO Anri Kivimäki explained that sports teams are already recording a lot of video of their matches (both for broadcast and internal use). Coaches then go through the video manually, or they send it off for someone else to create clips of the highlights, a process that can take a day or more.

The startup spun out of the University of Oulu in Finland, with computer vision technology that can look at video and identify different types of plays — and eventually, individual players.

That means it can digest footage and automatically create a timeline for the coach that allows them to jump to the most relevant moments in a game (and they get to specify what’s relevant to them). In the future, a coach might be able to share all the clips featuring a single player with that player.

[gallery ids="1574043,1574141,1574142,1574139,1574138"]

AISpotter charges 100 euro per video analyzed. Kivimäki said it usually takes about an hour to create the analysis, though the platform can also take advantage of other data about the game to shave that time down to 10 minutes.

She acknowledged that there are products trying to automate parts of the coaching process, but she said those products are focused on statistics, not videos. “Some companies produce 60 pages of numerical data, but for coaches and for players, it’s not that helpful.”

The company is currently aiming to work with football (soccer) and ice hockey teams: Football because it’s the most popular sport in the world, hockey because in Finland, “We are good in hockey.” Kivimäki isn’t necessarily looking to sign up top-tier teams (though AISpotter would happily work with them), but rather professional teams in the levels below the Premier League.

And again, the teams are usually recording the needed footage already, but if necessary, AI Spotter can also install the necessary (off-the-shelf) cameras.

Ultimately, Kivimäki and her CTO Sami Huttunen said the technology could be applied to other sports. And it could be used in areas outside coaching, like fan engagement, betting and scouting.