ALL PROJECTS / 2019

Fencing Visualized

Fencing Visualized

Year: 2019

Client: Rhizomatiks / Dentsu Lab Tokyo, Tokyo, Japan

Technology: Python, YOLOv3, convolutional neural networks, multi-camera 3D tracking, AR trajectory visualization

Reference Links

Video on YouTube

External Links

"Fencing tracking and visualization system" for Rhizomatiks "Fencing Visualized at the World Fencing League (LA)" for Rhizomatiks "Fencing Visualized Data Archive" for Rhizomatiks "Fencing Visualized Data Archive"

Description

Software engineering for Rhizomatiks’ markerless sword-tip tracking system, using deep learning to turn high-speed fencing action into real-time 3D trajectories.

For the 2019 All Japan Fencing Championship and related World Cup work, IYOIYO contributed to the pipeline that detected sword tips from camera imagery, estimated 3D paths, and prepared data for AR-style broadcast visualization. The system evolved from marker-based prototypes into a deep-learning approach capable of tracking motion that is difficult to follow by eye.