A modified fuzzy c-means algorithm for breast tissue density segmentation in mammograms

Zhili Chen,Reyer Zwiggelaar

DOI: https://doi.org/10.1109/ITAB.2010.5687751

2010-01-01

Abstract:The fuzzy cmeans (FCM) algorithm has been applied in a variety of medical image segmentation applications. The conventional FCM algorithm uses the greylevel information at a single pixel as the feature space and this contains no spatial contextual information, which makes it very sensitive to noise and intensity inhomogeneities. Recently, some modified FCM algorithms with spatial constraints have been published. However, these have individual disadvantages and are not robust enough with different types of noise. In this paper, we propose a modified FCM algorithm incorporating local spatial and intensity information based on an adaptive local window filter whose weighting coefficients differentiate the neighbouring pixels within the local window. Fast clustering is afterwards performed on the intensity histogram of the filtered image. To demonstrate the robustness and insensitivity to noise of the proposed algorithm, it is extensively tested using synthetic images corrupted by a variety of noise. The experimental results are quantitatively evaluated and compared. This algorithm is then applied to mammographic images for breast tissue density segmentation. The segmentation results indicate its effectiveness to the presence of intensity inhomogeneities in mammograms from different density categories. © 2010 IEEE.