HOS-based pattern classification and clustering for effective moving object detection

This paper presents a novel approach for detecting moving objects in image sequences through a distribution based model. The proposed methodology leverages higher-order statistics (HOS) within a clustering framework to enhance the accuracy and effectiveness of moving object detection. The proposed method effectively combines statistical knowledge about the class of moving objects with motion information. By utilizing HOS-derived data from sample images, the unknown distribution of object image patterns is approximated. Our proposed algorithm uses an HOS-based decision measure which is derived from a series expansion of the multivariate probability density function in terms of the multivariate Gaussian and the Hermite polynomial. The clustering process, guided by HOS, enhances the decision-making process by enabling a higher-order closeness measure to accurately classify test clusters as foreground or background.