Fast and consistent images areas recognition using an Improved Shuffled Frog Leaping Algorithm

Abstract

Improved Shuffled Frog Leaping Algorithm (IMSFLA) is a novel optimal algorithm of automatic recognition of areas of greyscale-based images. It is as an efficient improvement of the original Shuffled Frog Leaping Algorithm (SFLA). SFLA is a newly developed evolutionary algorithm with good global search capability. In this new paradigm, we propose a new fitness function. It is computationally simple and assists to quickly discover the adequate threshold in a continuous range of grey levels. And more, the new method is enhanced by the cloning of the fitter particles at the expense of the worst particles. The performance of IMSFLA is evaluated towards an Artificial Bee Colony algorithm (ABC) based method, two Genetic algorithm (GA) based method and Artificial Fish-Swarm (AFS) based method for many benchmark images and IMSFLA outperforms these algorithms.