Evolution of optimised structuring element sequences for high-speed embedded vision applications

Abstract

This paper presents a variable length steady-state genetic algorithm with a novel weighted diversity management scheme for the optimisation of structuring element sequences for a high-speed vision application. The weighted diversity measure reflects the hierarchical structure of the genome and improves the diversity and termination rate of the algorithm. The results show that the algorithm is able to take advantage of the large instruction set of the processor and reductions in sequence lengths of greater than a half are achieved when compared to previous implementations.