Fig 1.
A demonstration of traditional recognition and detection algorithms of table tennis.
Fig 2.
Flowchart of the tactical indicator in an automatic detection system.
Fig 3.
DDPG algorithm based on deep reinforcement leaning.
Fig 4.
Structure diagram of convolution neural target tracking in table tennis.
Fig 5.
Structure of LSTM unit.
Fig 6.
Structure of table tennis trajectory prediction.
Fig 7.
Real-time data acquisition system.
Fig 8.
Feature extraction results based on deep reinforcement leaning.
Note: common feature extraction algorithms include SIFT (Scale Invariant Feature), SURF (Speeded Up Robust Feature), ORB (ORiented Brief), FAST, and DPG (Deterministic Policy Gradient).
Fig 9.
Target tracking results based on CNN.
Note: common target tracking algorithms include CF (Collaborative Filtering), DPN (Deformable Parts Model), SVM (Support Vector Machine), RPN (RPN+Fast RCNN), and YOLO (You Only Look Once).
Fig 10.
Trajectory prediction results based on LSTM network.
Note: common trajectory prediction algorithms include ID3 (Iterative Dichotomiser 3), BP (Backpropagation), RNN, ARIMA (Autoregressive Integrated Moving Average model), and Kalman gain.
Fig 11.
Overall performance evaluation results of the model.
Note: TM represents the traditional detection method of table tennis tactical indicators, and DCNN-LSTM represents the fusion algorithm model in DDPG+CNN+LSTM.