BetConstruct’s Head of Science Dmytro Fedyukov talked to iGaming Business magazine revealing details about AJNA, an AI-powered tool for on-field observations.
There are many benefits to increasing the accuracy of data scouting in sports betting. As well as increasing the trust and improving the experience and gamification for players, collecting more technical data opens up new and unique markets to bet on.
However, our capabilities of doing this manually are limited; technology plays a fundamental role in taking data scouting to the next level. One company striving to incorporate innovative technology into its data scouting solutions is BetConstruct. Its new product, AJNA, uses a combination of augmented live video streaming and artificial intelligence to produce new standards of data collection and analysis.
Dmytro Fedyukov, head of data science at BetConstruct, explains further: “This development was all about collecting statistical information in real-time using a combination of hardware and software. Using quite simple cameras, we can extract statistics data for all players on the field.” The technology was built with three aims:
• To develop statistical information for teams and coaches for tactical analysis
• To create augmented information and add more information to live streams
• To collect individual statistics and create additional opportunities in sports betting
Integrating AI technology
Fedyukov breaks down how the AI component of AJNA works alongside the live streaming: “We have several phases in our process. The first is input stream processing, where we do un-distortion, remaps, stitching, rotates for video frames, and identify scheme and field corner coordinates. In other words we automatically define full-field geometry from the video stream. “In the next phase, we use the player detection module to track players, dividing all people on the game field into team and referee categories. With this data we are able to track in a robust way and store personal tracks as geo-spatial event logs for each player. On top of that, we run event detection models that allow us to extract game semantic information such as goal likelihood.”
With these increased capabilities come new opportunities for sports betting verticals. The tracker then takes this information and collects the statistical data, which can be used to predict various metrics such as who has the ball at which point, whereabouts on the field they are and where other players are in relation to them. The technology can predict goal likelihood in real-time. The combination of AI and live streaming offers extended personal tracking like player speed or distance run per game, which can be converted into a number of new markets to bet on.
Of the product, Vigen Badalyan, Founder and CEO of BetConstruct says: “BetConstruct is about providing viability and convenience. We come up with creative and enlightening solutions such as the AJNA live tracking device which serves as a third eye on the gaming field, seeing way more than the eyes of a scout.”
Currently, AJNA provides full live on-field coverage of football, with a view to extending this to more sports in the future.