Article: Chaotic clonal selection optimisation for multi-threshold segmentation Journal: International Journal of Signal and Imaging Systems Engineering (IJSISE) 2015 Vol.8 No.5 pp.298 - 315 Abstract: As the computational complexity of multi-threshold segmentation methods grow exponentially, a number of bio-inspired algorithms have been emerging as the potential solutions due to nature of their linear growth. However, the use of traditional pseudo-random number leads to instability of the solution. This study proposes new Clonal Selection Algorithm (ClonalG) that uses deterministic chaotic dynamical system to replace the use of pseudo-random number. We use logistic map to generate the chaotic number with uniform distribution. The chaotic number is used to substitute the pseudo-random number for the population initialisation, affinity maturation and new antibodies generation. The algorithm behaves stable in terms of the final solution. The proposed algorithm was evaluated and used for multiple image segmentation on nine standard test images. The results confirmed that the proposed algorithm was more effective and stable in comparison to conventional ClonalG and Particle Swarm Optimisation (PSO). Inderscience Publishers - linking academia, business and industry through research

Title: Chaotic clonal selection optimisation for multi-threshold segmentation

Authors: Habibullah Akbar; Nanna Suryana; Shahrin Sahib

Addresses: Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, 76100 Durian Tunggal, Melaka, Malaysia ' Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, 76100 Durian Tunggal, Melaka, Malaysia ' Faculty of Information and Communication Technology, Universiti Teknikal Malaysia Melaka, Hang Tuah Jaya, 76100 Durian Tunggal, Melaka, Malaysia

Abstract: As the computational complexity of multi-threshold segmentation methods grow exponentially, a number of bio-inspired algorithms have been emerging as the potential solutions due to nature of their linear growth. However, the use of traditional pseudo-random number leads to instability of the solution. This study proposes new Clonal Selection Algorithm (ClonalG) that uses deterministic chaotic dynamical system to replace the use of pseudo-random number. We use logistic map to generate the chaotic number with uniform distribution. The chaotic number is used to substitute the pseudo-random number for the population initialisation, affinity maturation and new antibodies generation. The algorithm behaves stable in terms of the final solution. The proposed algorithm was evaluated and used for multiple image segmentation on nine standard test images. The results confirmed that the proposed algorithm was more effective and stable in comparison to conventional ClonalG and Particle Swarm Optimisation (PSO).

Keywords: chaos theory; clonal selection; multilevel thresholding; bio-inspired computation; PSO; particle swarm optimisation; Otsu criterion; chaotic dynamical systems; pseudo-random numbers; logistic maps; image segmentation; multiple images.

DOI: 10.1504/IJSISE.2015.071953

International Journal of Signal and Imaging Systems Engineering, 2015 Vol.8 No.5, pp.298 - 315

Accepted: 20 Sep 2013
Published online: 25 Sep 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Chaotic clonal selection optimisation for multi-threshold segmentation

Keep up-to-date