Modified Bias Field Fuzzy C-Means for Effective Segmentation of Brain MRI

In recent day, segmentation of brain Magnetic resonance Image (MRI) with bias field correction is challenging and unavoidable in high magnetic imaging. The brain MRI is affected by bias field that causes the undesired effect of quantitative image analysis. The removal of bias field distortion is useful in segmenting medical images for proper study of medical images. In this paper, we propose three new Fuzzy c-Means (FCM) algorithms namely Robust Gaussian based Weighted Bias Field FCM [RGBFCM], Spatial constraint Gaussian based bias field FCM [GBFCM_S], Novel Penalized Gaussian based Bias field FCM [NPGBFCM] in order to remove bias field distortion and to obtain well segmentation result. The proposed methods are capable to deal with the intensity in-homogeneities and noisy image effectively. Further, to reduce the number of iterations, the proposed algorithms initialize the centroid using dist-max initialization algorithm before the execution of algorithms iteratively. To show the performance of proposed methods, this paper applies them to segmentation of brain MRIs and compares the results of our proposed methods with other reported methods. The segmentation accuracy of proposed method is validated by using Silhouette method. The experimental results on real T1-T2 weighted and simulated brain MRIs show that our methods are superior in providing better segmentation results than standard fuzzy c-means based algorithms.