A Novel Sports Video Background Segmentation Algorithm Based on Graph Partition

Abstract

In this paper, we propose a novel sports video background segmentation algorithm based on graph partition, and this topic is very important in multimedia processing and retrieval. Firstly, we illustrate the graph partition problem, which aims to divide a graph into several parts under some constraints. Furthermore, a cost function in the graph partition problem is defined. Secondly, sports video background segmentation algorithm based on graph partition is provided. Particularly, the main innovation of this paper lies in that we introduce the graph partition algorithm, and we convert the image segmentation problem to an energy minimization problem. In the end, experiments are conducted to make performance evaluation, and the Olympic Sports Dataset is used. Experimental results show that compared with existing methods, our algorithm can achieve high quality of sports video background segmentation.