Image Segmentation of the Continuous Action of Spiking in Volleyball Based on Spatial Neighborhood Information

Abstract

In real scenes of sports competitions, interferences like light intensity and camera jitter make it difficult to accurately segment and identify the action information in the complex continuous actions of athletes. Spiking, a common action in volleyball, can be divided into three closely correlated phases: landing, buffering, and stretching. This paper explores the image segmentation of the continuous action of spiking in volleyball based on spatial neighborhood information. Specifically, the authors detailed how to acquire the neighborhood correlation images of volleyball players spiking the ball. The framework of the image segmentation model was presented, and the extraction method of weighted spatial neighborhood information was expounded. Next, the fuzzy c-means (FCM) clustering was optimized to effectively segment the video images on the continuous action of spiking in volleyball. The proposed algorithm was proved valid through experiments.

Fuzzy C-Means Clustering for Tennis Multivariate Time-Series

“…Many studies focus on image or video processing using clustering in sport science. In Yang (2022) the segmentation on the continuous action of spiking in volleyball based on spatial neighborhood information, using C-Means was presented. The analyzed move was divided into the following phases: Landing, buffering, and stretching.…”