Deep learning model analyzes table tennis player skills from smart racket data
TTNet: Multi-Task Deep Learning for Table Tennis Player Analysis with Smart Racket
Machine Learning
Summary
Analyzing how people play table tennis can be complicated, but new smart rackets collect detailed swing data to help. The authors created a deep learning model called TTNet that can look at this data and guess things about the player, like their gender, which hand they use, how experienced they are, and their skill level. This helps make better and more precise assessments of player skills. Their smart method performed well in a competition, showing promise for using AI to study sports moves.
What this means in practice
- •For sports equipment manufacturers: Develop smart sports gear that can analyze player attributes and skills in real-time using sensor data like six-axis measurements from rackets.$Commercial implications: Enables products like smart rackets that provide players and coaches with actionable insights about performance and skill levels.
- •For sports coaching teams: Integrate advanced player attribute prediction models to enhance training programs by accurately tracking experience and skill levels from sensor data.
Authors
Ko-Hsun Chen, Xiang-Wei Ke, Hsien-Cheng Huang, Shang-Kuan Chen
Abstract
The AI CUP 2025 Precise Analysis of Table Tennis Smart Racket Data Competition introduced smart table tennis rackets that collect extensive player swing data, enabling research on table tennis big data. These data support in-depth analysis of players' return techniques and swing-force consistency, improving the accuracy of player skill assessment. This study focuses on six-axis sensor data collected by smart table tennis rackets and proposes TTNet, a novel deep learning model with multitask learning capabilities, to advance table tennis data analysis and related applications. TTNet combines convolutional neural networks (CNNs), residual networks (ResNet), and self-attention mechanisms to simultaneously predict four player attributes: gender, playing hand, years of experience, and skill level. We adopt a two-stage training strategy that incorporates data augmentation and task-specific loss functions to improve generalization on imbalanced data. Our approach achieved second place on the official competition leaderboard.