Research on Basketball Shooting Action Based on Image Feature Extraction and Machine Learning

In modern sports training, collecting and analyzing basketball player's posture data is of great significance for improving the scientific of the coach's training plan and improving the athlete's training effect. The existing basketball action recognition technology has many challenges such as low efficiency and high error rate. In order to effectively identify the basketball player's sports posture and improve the athlete's training effect, this paper proposes a basketball shooting gesture recognition method based on image feature extraction and machine learning. First of all, the action posture data of basketball players is collected by image feature extraction method, and multi-dimensional motion posture features are extracted from time domain and frequency domain. Then, through the method of feature selection and Gaussian hidden variables, the accurate classification and recognition of basketball shooting gestures are realized. The actual case analysis and the assessment of shooting action recognition effect show the superiority of the achieved basketball shooting action recognition technology. This method can provide scientific reference and basis for the development of modern basketball training.