Evolutionary Computation for Figure-Ground Separat ioin

The problem of figure-ground separation i s modeled as one of energy minimizat ion using the Ising sy s t em model f r o m quantum physics. T h e Ising syst e m model f o r the figure-ground separation problem makes explicat the definition of shape in t e r m s of attributes such as cocircularity, smoothness, proximity and contrast and i s based on the formulat ion of an energy funct ion that incorporates pairwise interactions between local image features in the f o r m of edgels. The paper explores a class of stochastic opt imizat ion techniques based on evolutionary algorithms i n the context of figure-ground separation using the Ising sy s t em model. Experimental results on synthetic edgel maps and edgel maps derived f r o m gray scale images are presented.