.Optimization of Table Tennis Swing Action Supported by the Temporal Convolutional Network Algorithm in Deep Learning

To improve the swing of table tennis more accurately, an improved temporal convolutional network (TCN) algorithm is proposed to capture the temporal relationship and spatial characteristics in the swing action. Firstly, the basic structure of table tennis swing action recognition based on TCN has been studied, and algorithm improvements have been made to recognize table tennis swing action. In the algorithm improvement stage, the activation function is adjusted by replacing the traditional Rectified Linear Unit (ReLU) with Leaky ReLU, effectively avoiding the problem of gradient vanishing and better capturing the temporal relationships and spatial characteristics in table tennis swing actions. Secondly, the network structure is optimized by replacing the fully connected layer with the global average pooling layer to reduce the complexity and computational burden of the model. Finally, the residual structure in the network is fine-tuned to enhance the model's adaptability to swing action features. In the experimental stage, this study evaluates the model using the OpenTTGames dataset, which contains 55582 data samples, and divides the training and testing sets in a 3:2 ratio. The results reveal that the proposed improved TCN has achieved significant results in table tennis swing action recognition, with an algorithm recognition accuracy of 99.43%, and a recall, accuracy, and F1 value of 99.00%. The recognition accuracy of this algorithm is 10.57%, 3.65%, and 2.70% higher than that of TCN, Long Short-Term Memory (LSTM), and Convolutional Neural Network-LSTM (CNN-LSTM) algorithms. The recognition performance of the model is successfully improved, providing strong support for the technical training and competitive performance of table tennis players. This research result has important practical significance for technological improvement in sports and the application of artificial intelligence in sports training.