Abstract:In recent years, developments in the fields of computer vision and artificial intelligence have created new opportunities for studying sports performance. With advancements in computer vision and artificial intelligence, it is now possible to use massive volumes of video data to get deeper insights into sports dynamics, especially in precision-based video sports like table tennis. To develop skills, a thorough examination of player movements is required. With the development of vision-based human posture recognition, computers can function like humans and derive intelligent judgments from outside data. This paper presents a novel method that uses graph convolutional neural networks (GCNNs) to detect and identify improper postures and movements in table tennis players. In addition to conventional methods, the proposed method dissects the human skeleton into finely detailed head, trunk, and leg features. Deep-level features that offer a more comprehensive understanding of athlete movements are extracted after feeding these features into the network. The softmax classifier combines these features to produce the final recognition result. The effectiveness of this approach has been evaluated through extensive experimentation. The GCN model performs remarkably well, with accuracy rates of 86.4% on the NTU-RGB + D dataset and 79.1% on the COCO dataset. This accomplishment is significant because the model consistently yields accurate detection results, even in complex and occlusion-filled scenes. In comparison tests, the proposed model performs better than GraphSAGE, GAT, ChebNet, GIN, and GC-LSTM in identifying relevant body part movements in table tennis players while ignoring irrelevant ones. According to proposed experiments, this enhances the recognition of improper postures in most table tennis actions.

Robust Stroke Recognition via Vision and IMU in Robotic Table Tennis

Markerless Racket Pose Detection and Stroke Classification Based on Stereo Vision for Table Tennis Robots

A Learning Framework Towards Real-Time Detection and Localization of a Ball for Robotic Table Tennis System

Real-time vision system for a Ping-Pong robot

A Table Tennis Robot System Using an Industrial KUKA Robot Arm

Spin Observation and Trajectory Prediction of a Ping-Pong Ball

A multi-modal table tennis robot system

Balance motion generation for a humanoid robot playing table tennis.

Real-time Accurate Ball Trajectory Estimation with "asynchronous" Stereo Camera System for Humanoid Ping-Pong Robot.

Table Tennis Robot Based on High Speed Vision

Table Tennis Stroke Detection and Recognition Using Ball Trajectory Data

A Model-free Approach to Stroke Learning for Robotic Table Tennis

Spin Detection in Robotic Table Tennis

Binocular Vision-Based Recognition Method for Table Tennis Motion Trajectory

A Robust Vision Module for Humanoid Robotic Ping-Pong Game

Computational Analysis of Table Tennis Matches from Real-Time Videos Using Deep Learning.

Table Tennis Robot With Stereo Vision And Humanoid Manipulator Ii: Visual Measurement Of Motion-Blurred Ball

A novel approach for automatic detection and identification of inappropriate postures and movements of table tennis players

Dexterous Robot Arm for Table Tennis and Hitting Strategy

Optimal stroke learning with policy gradient approach for robotic table tennis

Robotic Table Tennis: A Case Study into a High Speed Learning System