Creative Commons source reading · Data quality and algorithm evaluation
Applying Deep Learning to Image and IMU Sensor Data for Real-Time Fatigue Recognition in Table Tennis Players
Sheu YH; Tai TH; Chen TY; Huang CY; Wu SK
Originally published in Sensors (Basel) · 2026-10-09
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
AbstractFatigue during table tennis training can impair performance and increase the risk of injury. This study proposes a real-time fatigue recognition system that integrates an embedded edge-AI racket with a camera for synchronized multimodal data acquisition. On the racket side, a neural network was deployed on an edge-AI microcontroller unit (MCU) to enable on-device inference, while the test data were transmitted to a computer for multimodal feature fusion. Given the continuous nature of fatigue development, the data were divided into three stages: non-fatigue, transition, and fatigue to facilitate progressive evaluation. In the binary classification task, which distinguished between the two extreme states, the multimodal hybrid model achieved an accuracy of 99%. When the transition stage was incorporated into a three-class…
4. Conclusions
This study makes three main conclusions. First, it demonstrates a dual-modality fatigue-monitoring framework combining an embedded IMU racket with vision-based pose estimation, fused through weighted soft-voting decision fusion, that is feasible and computationally lightweight for use during table tennis training. Second, by dividing fatigue accumulation into three interpretable stages, the framework distinguished the two most extreme states with 99% accuracy in a binary formulation and reached 84–87% accuracy in a three-class formulation that improved to 96% after multimodal fusion, indicating that multimodal fusion may particularly aid discrimination during the transition between non-fatigue and fatigue, where single-modality models appear more prone to misclassification. Third, embedding the circuit board within the racket handle…
Editorial reading guide
- What data created and tested the model, and was evaluation independent?
- What is the cost of a false positive, false negative or out-of-distribution decision?
