Introduction

Abstract

One of the most challenging issues in computer vision is image segmentation. The reason lies on the information it can provide about the elements in the scene from the automatic image division based on pixel similarities. Therefore, what makes a pixel interesting depends on the object's features to be considered. Thus, due to segmentation of countless applications, a wide range of solutions have been proposed and tested by the scientific community during the previous years. However, considering motion as a primary cue for target detection, background subtraction (BS) methods are commonly used. In this chapter, we overview the method in general terms as well as its different variants with the aim to analyze the problems remaining to be solved.