An optimized fuzzy clustering algorithm for brain magnetic resonance image segmentation

Abstract

Magnetic Resonance Imaging (MRI) technique is an important adjunct to the clinical diagnosis of brain disease with its advantages of non-invasive and noninvasive. The precise segmentation of brain MR images is a significant issue for biomedical research and clinical applications. However, there are various defects in brain MR images, such as gray irregularities, noise, and low contrast, reducing the accuracy of the brain MR images segmentation. In this paper, we propose an optimization solution for a fuzzy clustering algorithm and apply it in brain MR image segmentation. A fuzzy C-means clustering algorithm is proposed based an anisotropic weight model. By introducing the new neighborhood weight calculation method, each point has the weight of anisotropy, effectively overcome the influence of noise on the image segmentation. In addition, two kinds of brain MR image segmentation models are proposed to address the low contrast defects in brain MR images. The local Gaussian probability model is included in the objective function of fuzzy clustering, and a clustering segmentation algorithm of fuzzy local Gaussian probability model is proposed. A clustering segmentation algorithm of adaptive scale fuzzy local Gaussian probability model is proposed. The neighborhood scale corresponding to each pixel in the image is automatically estimated, which improves the robustness of the model and achieves the purpose of precise segmentation.