Medical image segmentation needs higher segmentation accuracy due to the occurrence of noises in the images. Fuzzy sets theories are able to handle the vagueness and uncertainty through membership functions in image segmentation. Rough sets theory (RST) is to deal with uncertainty and incompleteness. It focuses on the feature selection based on classification. In real applications, it may be impossible to obtain complete information of a given pattern set due to artifacts. Uncertain information will cause lacking of information for a pattern set in various recognition and classification algorithms. This paper proposes a method to segment the magnetic resonance images with and without noises powerfully. The proposed method uses the intuitionistic fuzzy c-means algorithm for segmenting cerebro spinal fluid (CSF), white matter (WM) and gray matter (GM) tissues in the MRI. Intuitionistic fuzzy image representations are done by using non-membership value, hesitation along with the membership value for the MR image. The membership value and nonmembership value have been obtained using fuzzy trapezoidal membership and fuzzy complement function respectively. Further, intuitionistic fuzzy roughness measures and fuzzy c-means clustering determines the initial cluster centroids by considering lower and upper approximation and updates the euclidean distance between the pixels. The proposed method have been implemented and analyzed with performance metrics for the synthetic and real MR images. Experimental results reveal a superior degree of segmentation on both synthetic and real MR images compared to existing methods.