Image Segmentation by a Robust Clustering Algorithm Using Gaussian Estimator
This paper presents a novel clustering-based image segmentation method, which incorporates the features of robust statistics. To overcome the sensitivity to noise and outliers in fuzzy clustering, a simple but efficient M-estimator, Gaussian estimator, has been introduced to clustering analysis as weight or membership function. When applied to image analysis, the proposed Robust Gaussian Clustering/Segmentation (RGCS) algorithm exhibits more reasonable pixel classification and noise suppression performance with respect to FCM. Moreover, by using a uniform resolution parameter scheme, this method avoids producing coincident clusters, which occurs often in possibilistic clustering segmentation. Experiments on both the synthetic data and real image demonstrate the validity and power of the proposed algorithm.
Unable to display preview. Download preview PDF.
- 1.Rosenfeld, A., Kak, A.C.: Digital Picture processing, 2nd edn. Academic Press, New York (1982)Google Scholar
- 8.Goodall, C.: M-estimator of location: An outline of the theory. In: Hoaglin, D.C., Mosteller, F., Tukey, J.W. (eds.) Understanding Robust and Exploratory Data Analysis, New York, pp. 339–403 (1983)Google Scholar