T-Fort:A Tree-Based Fast-Moving Object Trajectory Tracking Algorithm for Table Tennis
Tianjian Zou, Jiangning Wei, Hao Zhang, Zechen Jin, Yang Yu, Jun Liu · 2023
Table tennis is a popular sport with high popularity around the world. A key technology in table tennis analysis system is reconstructing the trajectory of table tennis by tracking the fast-moving from images or videos. As the ball is always barely visible and too tiny in the image or video to be detected directly by detection models such as YOLO and Fast- RCNN, researchers usually track it by analyzing their moving patterns. In this article, we propose a tree-based algorithm named T-FORT to tracking the ball and its trajectory. Specifically, we consider all the possibly in a tree-framework, and select the target by some known parameters. We then evaluate the proposed algorithm in a table tennis video dataset, the results show that our method is more precise than algorithm before, and can be widely used.