Hello !
My name is Ziyao Sun, but I’m better known as Josh, and I am currently a high school senior. I began fencing at the age of 9 and switched from Épée to Sabre after moving to North Carolina. Since 2018, I have been training at Forge Fencing Club in Durham. In addition, I hold a regional sabre referee license.
My Idea :
The challenge of performance analysis in fencing videos stems from the rapid pace of the sport, making it difficult for computers to accurately detect such fast movements. To address this, I propose the following approach: First, use AI to detect slower moments, such as the En Garde position. Once detected, the AI can split the video into individual clips of touches. By allowing users to label these clips with specific fencing actions, the AI can then perform statistical analysis. Using this approach, the challenges become more manageable.
The Needs,
After fencing for so long, I’ve noticed an issue in the sport: while fencing is growing, our analysis tools aren’t keeping up. Right now, most fencers rely on parents or friends to record their bouts, and those videos are then sent to them or their coaches for review. Observations are made manually, which works, but it’s super inefficient—and there’s no system to track or keep records of the action analysis.
Thank You !
To my coaches, Coach Jen, Coach Jeff, Coach Stephen and Coach Cathleen for helping me understand the needs of coaches and fencers;
​
To all my teammates, especially Tyson and Nora, for sharing their fencing videos with me to use as training data;
​
To my technical advisors, Hanlong Yang and Yuanqi He, for assisting me during the challenging development process.
​
To my teachers, Ms. McKenzie and Ms. Ramey, for validating the approach and guiding me through the entire software development process.
.png)