Development of fuzzy clustering based unsupervised scheme for medical image segmentation using HMRF model

In this paper, the problem of medical image segmentation is addressed in an unsupervised framework. We propose a novel method considering the hidden Markov random field model (HMRF) to model the image class labels, which takes into account the mutual influences of neighboring sites formulated on the basis of fuzzy clustering principle. The model parameters, number of class labels and the image labels are assumed to be unknown. Here an attempt has been made to incorporate the benefits of HMRF model into the benefits of fuzzy clustering procedure. To combine the spatial coherency modeling capabilities of the HMRF model and the enhanced flexibility obtained by fuzzy c-means (FCM) algorithm, fuzzy clustering expectation maximization (FCEM) algorithm is proposed. The initial model parameters are assumed arbitrarily unlike existing methods. Both model parameters as well as class labels of medical images are estimated recursively using proposed algorithm until the model parameters converge to the optimal ones. The proposed HMRF-FCEM segmentation scheme is validated with various noisy medical images. We experimentally demonstrate the superiority of the proposed approach over the existing HMRF-EM framework applied to medical image segmentation. The proposed scheme does not depend on the initial choice of model parameters and can be applied for automatic medical image analysis.