ABSTRACT
This chapter discusses mathematical models of images that can be represented with graphs, known as graphical models. These models are powerful because the graphs provide a powerful mechanism for describing and understanding the structure of the models. The chapter reviews how these models are typically defined and used in computer vision systems. One of the steps in creating a computer vision system is the establishment of the overall computational paradigm that will be used to compute the final solution. Undirected graphical models are a natural tool for modeling patches of pixels in low-level image models. Undirected graphical models are a natural tool for modeling patches of pixels in low-level image models. Graphical models are a powerful tool for building mathematical models of images and scenes.
