Classical Mechanics Optimization for Image Segmentation
In this work, we focus on image segmentation by simulating the natural phenomenon of the bodies moving through space. For this, a subset of image pixels is regularly selected as planets and the rest as satellites. The attraction force is defined by Newton’s third law (gravitational interaction) according to the distance and color similarity. In the first phase of the algorithm, we seek an equilibrium state of the earth-moon system in order to achieve the second phase, in which we search an equilibrium state of the earth-apple system. As a result of these two phases, bodies in space are constructed; they represent segments in the image. The objective of this simulation is to find and then extract the multiple segments from an image.
KeywordsImage segmentation Combinatorial optimization Artificial intelligence Metaheuristic Classical mechanics optimization
- 2.Rashedi, E., Nezamabadi, H., Saryazdi, S.: Gsa. A gravitational search algorithm. Information Sciences, 179(13): 2232–2248 (2009)Google Scholar
- 3.Amandeep, K., Charanjit, S., Amandeep, S.B.: SAR image segmentation based on hybrid PSOGSA optimisation algorithm, vol. 4, issue 9 (2014). ISSN 2248-9622Google Scholar
- 5.Dahiya, A., Dubey, R.B.: Survey of some multilevel thresolding techniques for medical imaging, vol. 3 issue 7 (2015). ISSN 2347-3878Google Scholar
- 6.Flores, J.J., López, R., Barrera, J.: Particle swarm optimization with gravitational interactions for multimodal and unimodal problems. In: Sidorov, G., Hernández Aguirre, A., Reyes Garc\’ıa, C.A. (eds.) MICAI 2010, Part II. LNCS, vol. 6438, pp. 361–370. Springer, Heidelberg (2010)CrossRefGoogle Scholar
- 7.Bichot, C.: Elaboration d’une nouvelle métaheuristique pour le partitionnement de graphe Doctoral thesis. The Polytechnic National Institute of Toulouse (2007)Google Scholar