Modelling image processing with discrete first-order swarms

Abstract

So farmost applications of swarmbehaviour in image analysis use swarms as models for optimisation tasks. In our paper, we follow a different philosophy and propose to exploit them as valuable tools for modelling image processing problems. To this end, we consider models of swarming that are individual-based and of first order. We show that a suitable adaptation of the potential forces allows us to model three classical image processing tasks: grey scale quantisation, contrast enhancement, and line detection. These proof-of-concept applications demonstrate that modelling image analysis tasks with swarms can be simple, intuitive, and highly flexible.