Abstract:This research investigates the recognition of basketball techniques actions through the implementation of three-dimensional (3D) Convolutional Neural Networks (CNNs), aiming to enhance the accurate and automated identification of various actions in basketball games. Initially, basketball action sequences are extracted from publicly available basketball action datasets, followed by data preprocessing, including image sampling, data augmentation, and label processing. Subsequently, a novel action recognition model is proposed, combining 3D convolutions and Long Short-Term Memory (LSTM) networks to model temporal features and capture the spatiotemporal relationships and temporal information of actions. This facilitates the facilitating automatic learning of the spatiotemporal features associated with basketball actions. The model's performance and robustness are further improved through the adoption of optimization algorithms, such as adaptive learning rate adjustment and regularization. The efficacy of the proposed method is verified through experiments conducted on three publicly available basketball action datasets: NTURGB + D, Basketball-Action-Dataset, and B3D Dataset. The results indicate that this approach achieves outstanding performance in basketball technique action recognition tasks across different datasets compared to two common traditional methods. Specifically, when compared to the frame difference-based method, this model exhibits a significant accuracy improvement of 15.1%. When compared to the optical flow-based method, this model demonstrates a substantial accuracy improvement of 12.4%. Moreover, this method showcases strong robustness, accurately recognizing actions under diverse lighting conditions and scenes, achieving an average accuracy of 93.1%. The research demonstrates that the method reported here effectively captures the spatiotemporal relationships of basketball actions, thereby providing reliable technical assessment tools for basketball coaches and players.

What problem does this paper attempt to address?

This paper aims to solve the problem of basketball action recognition by implementing a three-dimensional (3D) Convolutional Neural Network (CNN) to enhance accurate and automated recognition of various actions in basketball games. The research first extracts sequences of basketball actions from public basketball action datasets, and then performs data preprocessing, including image sampling, data augmentation, and label processing. Then, a novel recognition model combining 3D convolution and Long Short-Term Memory (LSTM) networks is proposed to model temporal features and capture the spatial and temporal relationships of actions. By adopting optimization algorithms such as adaptive learning rate adjustment and regularization, the performance and robustness of the model are further improved. The effectiveness of this method is validated on three public basketball action datasets, with significant improvements in recognition accuracy compared to traditional methods such as frame difference-based methods and optical flow-based methods. The model also demonstrates strong robustness under different lighting conditions and scenes. This research provides reliable technical and tactical assessment tools for basketball coaches and athletes.

Basketball technique action recognition using 3D convolutional neural networks