Passion Project: Creating an AI Referee for Fencing
“Inspirit AI was an essential first step in my AI studies.”
Before Inspirit, Jason had a conceptual idea of how AI worked, but he decided to participate in the AI Scholars program to learn how to translate that conceptual knowledge into practical coding know-how. With Inspirit, Jason was exposed to different types of AI models and learned enough about AI frameworks like Cv2 and Tensorflow to actually pursue projects after the program ended.
“Being introduced to Google Colab was also a life-saver.”
Bridging the Gap Between Passions: Sports and Tech
Jason got the idea for an AI referee for fencing when he was just starting to learn about AI.
“I knew from personal experience that refereeing was pretty repetitive and that there were tons of fencing videos on Youtube, so I realized that AI might work really well in fencing. It was also an exciting project because very few people have ever worked on an AI fencing referee so I couldn't just rely on a tutorial.”
Fencing is perhaps one of the hardest sports for spectators to watch. Not only do fencing actions occur in a fraction of a second, but fans also have to keep in mind the various rules of right of way. While interest in fencing peaked during the Tokyo Olympics thanks to Lee Kiefer’s amazing Gold finish, it quickly dissipated as people grew frustrated with the seemingly confusing point system. “Why doesn’t the first person to hit get the point?!?” People can’t enjoy fencing when they hardly understand what is going on.
Trend Analysis
A quick peek at Google trends for fencing clearly shows the effects of not being able to effectively explain the rules of fencing to newcomers, with interest peaking and quickly dissipating during the Tokyo Olympics.
From Inspiration to Reality: Bringing Allez Go to USA Fencing
In fencing, the concept of right of way dictates which fencer gets the point if both of them hit each other at the same. With his conceptual and technical understanding of AI, Jason was inspired to see if a computer could learn the rules of right of way.
Jason’s AI referee, Allez Go, steps in to simplify the spectator experience by revealing the complex world of right of way that is only typically seen by top coaches, referees, and fencers. Allez Go can be used during practice or small tournaments when there are no human referees available or for fencing livestreams. During livestreams, Allez Go displays which fencer it believes has right of way, making it easier for new spectators to see how right of way works. Jason is in the process of connecting with USA fencing (the national body for fencing) to see if they would be interested in using Allez Go.
Read more about Allez Go on our website.