A Novel Approach of Detection of Moving Objects in a Video

Abstract: In this paper, a novel approach is proposed for moving object detection and tracking. The main aim of this paper is to device a method by which moving objects can be detected in a video. A video stream is divided into several frames and our goal is achieved by identifying the moving portion which has changed over time. Here we used the temporal segmentation. In temporal segmentation, the absolute difference of two video image frames is segmented into changed and unchanged regions. We have divided the difference image frame into two clusters, one represents the changed region and the other represents the unchanged portion. Here changed region means moving object. We have achieved this by applying Genetic Algorithms (GA) using K-means clustering technique on the difference image frame. The proposed method is successfully tested over two video sequences.

Keywords: Genetic Algorithm, K-Means Algorithm, clustering, temporal segmentation, moving object detection, frame difference