Unsupervised Image Segmentation Using Automated Fuzzy c-Means
Authors:
- 1 October 2007
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. smc 8, 690-694
Abstract
An unsupervised fuzzy clustering technique, fuzzy c-means (FCM) clustering algorithm has been widely used in image segmentation. However, the conventional FCM algorithm must be estimated by expertise users to determine the cluster numbers. To overcome the limitation of FCM algorithm, an automated fuzzy c-mean (AFCM) algorithm is presented in this paper. The proposed algorithm initiates the first two centroids of clusters by a method based on Otsu algorithm and automatically determines the appropriate cluster number for image segmentation. The performance of the proposed technique has been tested with reference to conventional FCM. The experimental results demonstrate that AFCM can spontaneously estimate the appropriate number of clusters and its performance is faster convergence than the performance of the conventional FCM.Keywords
This publication has 6 references indexed in Scilit:
- Fuzzy c-means clustering with spatial information for image segmentationComputerized Medical Imaging and Graphics, 2006
- Image Segmentation Based on Model Selection2006
- Unsupervised Image Segmentation Using Penalized Fuzzy Clustering AlgorithmLecture Notes in Computer Science, 2005
- Pattern Recognition with Fuzzy Objective Function Algorithms1981
- A Threshold Selection Method from Gray-Level HistogramsIEEE Transactions on Systems, Man, and Cybernetics, 1979
- Fuzzy setsInformation and Control, 1965