An AI-based posture recognition system for analyzing the impact of core muscle strength on the forehand loop against backspin in table tennis

Abstract

Purpose

This study aims to evaluate the role of core strength in enhancing forehand loop performance in table tennis, leveraging an AI-driven posture recognition system developed using the MediaPipe algorithm.

Methods

Sixty participants, including 30 elite players and 30 school-level players, performed 30 forehand loops against standardized backspin balls delivered by a table tennis robot under controlled conditions. Motion data was captured at a 45° angle behind the players. Hip and knee angles were analysed using thresholds of 90° and 110° to measure core engagement. The AI system provided real-time motion analysis and stored results for further statistical evaluation.

Results

Elite players showed significantly higher core engagement compared to school-level players, achieving greater frequencies of shots meeting the predefined hip angle threshold of 110° (p < 0.001) and knee angle threshold of 90° (p = 0.004). The AI-driven posture recognition system demonstrated high accuracy in tracking motion and analysing biomechanics, effectively distinguishing skill levels. EP exhibited superior consistency and adaptability, with trends indicating improved technical efficiency across all performance metrics. These findings highlight the importance of core strength in enhancing forehand loop performance against backspin.

Conclusion

Core strength plays a pivotal role in optimizing forehand loop performance, particularly when countering backspin. This study highlights the potential of AI technologies, such as the MediaPipe-based system, to advance sports biomechanics and improve training methodologies in table tennis.