Exploring Fuzzy Clustering Objective Functions through Metaheuristic Optimizationfor Image Segmentation


Tezel Ö., Okuyucu O.

3. International Congress on Data Science and InformationTechnologies (INFTEC 2026), Lefkoşa, Kıbrıs (Kktc), 22 - 24 Ekim 2026, ss.36, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Lefkoşa
  • Basıldığı Ülke: Kıbrıs (Kktc)
  • Sayfa Sayıları: ss.36
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

Image segmentation is a fundamental process that involves identifying meaningful regions within animage and has a direct effect on the performance of image processing applications. Fuzzy clusteringmethods are widely used for image segmentation because they enable pixels to be assigned todifferent clusters with varying degrees of membership. However, these methods may face severalchallenges, such as sensitivity to the initial conditions, becoming trapped in local optima, anddifficulty in identifying optimal cluster centers in complex image structures. This study aims tosystematically assess the performance of different fuzzy clustering models optimized by swarmintelligence algorithms in image segmentation. In this context, different variants of the fuzzy c-means(FCM) clustering algorithm proposed in literature such as the outlier rejection fuzzy c-means(ORFCM), k-means integrated fuzzy c-means (KIFCM), intuitionistic fuzzy possibilistic c-means(IFCM) are considered. Unlike traditional fuzzy clustering approaches, the objective functions ofthese models are optimized using metaheuristic search mechanisms, with the aim of obtainingoptimal cluster centers. Within the experimental framework, the segmentation performance of thefuzzy clustering models optimized using metaheuristic methods is comparatively analyzed using testimages with different characteristics. To objectively assess the segmentation results, various imagequality and cluster validity metrics are implemented. Thus, this study presents a comprehensiveevaluation of the impact of objective functions on image segmentation performance by comparingdifferent fuzzy clustering models under the same metaheuristic algorithm.