Adaptively Regularized Kernel-Based Fuzzy C-Means Clustering Algorithm Using Particle Swarm Optimization for Medical Image Segmentation
| Autor: | Ibtissem Cherfa, Karim Mokrani, Anissa Mokraoui, Abdenour Mekhmoukh |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Kernel (image processing)
Analytical expressions Computer science 0202 electrical engineering electronic engineering information engineering Centroid Initialization Particle swarm optimization 020207 software engineering 02 engineering and technology Image segmentation Cluster analysis Algorithm Fuzzy logic |
| Zdroj: | SPA |
| DOI: | 10.23919/spa50552.2020.9241242 |
| Popis: | This paper is concerned with Magnetic Resonance (MR) brain image segmentation using Adaptively Regularized Kernel-Based Fuzzy C-Means (ARKFCM) clustering algorithm. However this algorithm is sensitive to the random initialization of the clusters’ centers and moreover its optimal solution can be trapped into a local rather than a global solution. To overcome these drawbacks, this paper proposes the Particle Swarm Optimization (PSO) strategy to compute the clusters’ centroids instead of using directly the derived analytic expression of the centroids given by the ARKFCM algorithm. Experimental results, carried out on MR brain images from the BrainWeb database, show that the revisited ARKFCM algorithm improves the performance of its original version. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|