Automatic Extraction of Moving Objects Using Distributed Genetic Algorithms

Abstract

This paper proposes an automatic segmentation method using genetic algorithms that can automatically extract the moving objects and track them. Each frame of the video sequence is decomposed into the objects forming the scene, which are tracked through a whole video sequence. The proposed method is evaluated on a number of different input data and is applied for car detection and people detection in the video surveillance system. The obtained results have confirmed the validity of the proposed method. © Springer-Verlag 2003.