Sequential blind image extraction based on particle swarm optimization

Lei Chen, Liyi Zhang, Yanju Guo, Ting Liu, Qiang Li

Abstract: A sequential blind image extraction method based on particle swarm optimization was proposed. Normalized fourth-order cumulant was used as cost function in the method and traditional gradient algorithm was replaced by particle swarm optimization algorithm for optimizing it. The extracted source image component was wiped off using deflation method and source images can be extracted sequentially. Simulation results show that source images can be effectively extracted from mixed images using this method. Then, a moving target detection method based on sequential blind image extraction method was proposed and the movement trajectory can be clearly observed from sequential images.