Title: Upper limb movement trajectory recognition of basketball players based on machine learning

Authors: Miao Wang; Limin Zhou; Cong Zhang

Addresses: Institute of Physical Education and Training, Harbin Sport University, Harbin 150001, China ' Institute of Physical Education and Training, Harbin Sport University, Harbin 150001, China ' Institute of Physical Education and Training, Harbin Sport University, Harbin 150001, China

Abstract: In order to overcome the problems of low accuracy of action classification and poor denoising effect of action signals in traditional motion trajectory recognition methods, the paper proposes a method for recognition of upper limb motion trajectories of basketball players based on machine learning. First, the characteristic change curve of the upper limb movement trajectory to extract the movement trajectory characteristic, is determined. Then, the wavelet transform method is used for signal denoising, and the support vector machine method in machine learning is used to design the movement trajectory recognition classifier to realise the movement trajectory recognition of the upper limbs of basketball players. The experimental results show that the method in this paper can effectively remove the noise in the upper-limb motion signal, improve the accuracy and recognition effect of upper-limb motion trajectory recognition, and has certain practical value in the recognition of basketball players' posture and motion.

Keywords: machine learning; support vector machine; wavelet transform; feature extraction.

DOI: 10.1504/IJICT.2023.132170

International Journal of Information and Communication Technology, 2023 Vol.23 No.1, pp.15 - 28

Received: 15 Mar 2021
Accepted: 25 May 2021
Published online: 12 Jul 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article