Leverage computer vision AI to identify key moments in our videos

Problem Statement

Our streaming product, for the viewers and the athletes, is all about capturing key moments in games. For the athlete, these moments represent the fruits of years of hard work and commitment, and for the viewers, these moments are the reason they subscribe to the streaming product in the first place.

Clips Beta is a fantastic MVP towards generating short-form videos that capture these key moments. It allows users to manually clip key moments and share them easily. The known limitation of this current approach is the time, effort, and initiative needed for users to sift through full games and generate these clips manually.

Proposed Solution

Use computer vision AI to automatically identify potential key moments on a player-by-player basis. This should involve some manual input to fine-tune start/stop times, but we should leverage AI to automate the heavy lifting of sifting through large volumes of video data to identify potential key moments.

The manual input could involve internal resources fine-tuning the start and stop time of clips, as well as “suggesting” potential clips for users to then go and create.

On the viewing side, we should allow users to easily view key moments for a particular team, and ideally, for the individual athlete they care most about.

In parallel to the above (i.e., an idea within the idea), we should work to enhance the viewing experience, hopefully with multi-angle and zoom capabilities, along with the ability to add effects and make these short-form videos more fun and engaging.

Unique Value Proposition

The core value of the streaming product is our ability to capture and showcase key moments to the people who care most about them. Viewers want to see their loved ones performing well in key moments and having an impact on the game. Athletes want to capture and relive these moments forever. Both proudly want to share these moments with others.

Quickly identifying as many of these key moments as possible and displaying them to users significantly increases the value of the streaming product. High quality short-form videos being shared widely would bring significant attention to the NFHS network product.