By Jay Jung ’27
Every second counts in the high-speed game of squash. Pragun Seth ’25 is transforming competitive tournament refereeing with his independent project, IntelliReferee, an AI-powered system that aims to automate decision-making during matches with an accuracy exceeding human referees.
The common disruptions caused by interference in squash games tend to be extremely distracting. These pauses can affect the flow of the game and lead to inconsistent refereeing decisions, impacting both the players’ performance and the audience’s visual experience. Seeing these conflicts and undesired pauses firsthand durisng both online matches and in-person junior tournaments, Pragun Seth ’25 was inspired to create an AI-based solution to improve refereeing accuracy. As an avid squash player since COVID-19, he utilized his passion for technology and entrepreneurship to develop IntelliReferee, giving back to the sport that had become a major aspect of his life. Seth explains, “I always noticed that matches were filled with stoppages due to interference debates, which felt unnatural for the sport. I wanted to create something meaningful that could address this issue using technology.”
Seth’s project was structured in three main stages: model design, data collection and training, and pilot testing. The first step was to understand the limitations of current refereeing practices in squash. He explored advanced computer vision models such as YOLO and TrackNet, which are capable of detecting interference situations based on player movements. The design process focused on customizing these models for recognizing interference in squash matches based on key metrics such as the positioning of players and movement paths.
The next following step involved data collection and processing. Seth established partnerships with the Professional Squash Association, World Squash Officiating, and the College Squash Association, to collect data for training the AI model. Such data included video footage from professional tournaments. By fine-tuning the AI model using diverse datasets, Seth aimed to ensure its reliability across different match settings, camera angles, and camera quality.
The last stage was the pilot testing of IntelliReferee. It was tested at major squash tournaments, such as the World Tour Finals and the Paris Open. During these events, the system effectively made real-time refereeing decisions, allowing the officials to observe its accuracy and make necessary adjustments.
Now, Seth plans to implement IntelliReferee at the U.S. Open and other major events, with a long-term goal to introduce a consumer version that can be used by amateur players via a smartphone app. Through commercializing IntelliReferee, Seth hopes to provide high-quality refereeing available to everyday players. Seth emphasizes, “I want to democratize access to professional-level refereeing, giving everyone the chance to experience fairer matches, whether they’re playing in a tournament or just with friends.”
He also plans to expand IntelliReferee’s capabilities beyond squash, with potential applications in other racquet sports like tennis and badminton. Through refining the AI algorithms, he aims to create a system that can adapt to different sports’ rules and dynamics. Seth is also considering collaborations with technology companies to further enhance IntelliReferee’s functionality and user experience.
Reflecting on his journey so far, Seth notes the broader implications of his work: “This project isn’t just about squash; it’s about how AI can bring meaningful change to sports and other areas. I hope IntelliReferee inspires more innovations in the field.”
Read more articles like this in our Fall 2024 Issue!
