Ant Colony for Locality Foraging in Image Enhancement

This chapter presents a spatial color algorithm called Termite Retinex and the problem of filtering locality in this family of algorithms for image enhancement, inspired by the human vision system. The algorithm we present is a recent implementation of Retinex with a colony of agents, which uses swarm intelligence to explore the image, determining in this way the locality of its filtering. This results in an unsupervised detail enhancement, dynamic range stretching, color correction, and high dynamic range tone rendering. In the chapter we describe the characteristic of glocality (\(\mathrm{glocal} = \mathrm{global} + \mathrm{local}\)) of image exploration, and after a description of the Retinex spatial color algorithm family, we present the Termite approach and discuss results.