Vietnam’s Pro Basketball League (VBA) is utilising an AI tool to automatically identify “key moments” in basketball games without manual intervention, which processes video feeds from the stadium, tags each frame with details like player involvement and actions, and ranks moments.
The ranked clips are then compiled into highlights of varying durations, tailored for platforms like TikTok, Facebook, and YouTube, enabling VBA to generate and publish more content, increasing the sport's popularity and engagement.
VBA’s chief operating officer, Tran Chu Sa, told iTnews Asia that the team spotted VideoVerse’s product – Magnifi, at a conference, and following a five-month experimentation period, the team decided to incorporate AI-driven video highlights of matches, video-on-demand, and social media functionalities for the 2023 season.
This resulted in a 76 percent increase in video views on Facebook and the amount of content that the team posted on YouTube went up by 177 percent, said Sa.
He added that it helps increase the number of mid-roll ads and social media presence, paving the way for better monetisation.
VideoVerse is an AI-focused video technology company and video-editing SaaS platform.
Earlier, VBA used to create highlights manually with the help of video editors and faced challenges in content production, experiencing delays in delivering game highlights to the audience.
Tran said the process of creating full game highlights used to take around 12 hours, causing the team to lose audience engagement. Now, with Magnifi, it takes 10 minutes, he added.
It has brought down the time of creating and publishing a moment from a game from two hours to nearly 15 seconds.
He added distributing content to TV networks was also cumbersome, as it required rendering heavy videos and facing difficulties in uploading online.
The tech working behind the platform
With automation in place, the VBA team can deliver fresh content to various channels on a nearly real-time basis, before others including spectators take ownership of the moment.
Magnifi’s senior vice president of revenue, Ritu Bain, said, the goal is to engage viewers across multiple platforms without the need for a large manual team.
Regarding technology, the product relies on computer vision, NLP (natural language processing), and ticker reading - which refers to the extraction and interpretation of information from on-screen tickers.
A ticker is a continuously updating text stream usually displayed at the bottom of a screen during a sports broadcast, providing real-time information such as scores, player names, and other relevant details.
According to Tran, the system triangulates information from player faces, jersey numbers, tickers, and commentary to create accurate tags for game moments. Patch checks and feedback mechanisms are in place, especially for fast-paced sports like basketball, ensuring the system's accuracy and reliability, he said.
Before each season begins, the team uses AI to analyse previous games, gaining insights into the teams, players, and overall dynamics.
Bain said the AI learns about the specificities of the league, including player attributes and team strategies, and this preparation helps ensure that when a new season starts, the AI already has a comprehensive understanding of the league's context.
Tran said VBA plans to focus on unifying its platform and use AI technology to leverage data effectively for commercial and monetisation purposes in the future.
It is looking to collaborate with advertising agencies to create a platform that streams content from Magnifi, ranging from highlights to live events to attract sponsors and advertisers interested in ways to engage with a young audience, he added.
