Title: Gesture recognition method for wearable sports devices based on sparse representation

Authors: Yuzhou Gao; Guoquan Ma

Addresses: Department of Physical Education, Lanzhou University of Technology, Lanzhou, Gansu Province, China ' Department of Physical Education, Lanzhou University of Technology, Lanzhou, Gansu Province, China

Abstract: In order to effectively solve the shortcomings of traditional gesture recognition methods for sports devices, such as high error rate and long recognition time, this paper designs a gesture recognition method for wearable sports devices based on sparse representation. A gesture model was constructed according to the gesture features of wearable sports devices, and the standard characteristic quantities of gestures were obtained. A sensor is then used to quickly collect a sample of the gesture. The sparse coefficient is obtained. Gestures were classified and processed according to the minimum residual value and sparse coefficient, so as to obtain accurate gesture recognition results of wearable sports devices. The simulation results show that the correct rate of gesture classification is always above 95.41%, the error rate of gesture recognition is always below 5.32%, and the recognition time is less than 0.8 s, which proves that the method has achieved good practical application effect.

Keywords: sparse representation; wearable sports devices; gesture recognition; characteristic; classification process.

DOI: 10.1504/IJPD.2023.129309

International Journal of Product Development, 2023 Vol.27 No.1/2, pp.41 - 53

Received: 04 Jun 2021
Accepted: 24 Nov 2021
Published online: 06 Mar 2023 *