This paper presents a novel algorithm, which uses intuitionistic fuzzy sets and rough set theory to segment the renal components in kidney MR images. A new membership function is proposed and then is used to obtain an intuitionistic fuzzy model of the image to compensate the inherent heterogeneity present among the different renal tissue classes. In addition, a new method, which uses Hamming distance is proposed to calculate the histon. The histon is then used to compute intuitionistic fuzzy roughness measure which yields optimum valley points for image segmentation. The proposed algorithm segments the kidney MR images into medulla, cortex, and blood vessels. The quantitative performance evaluation indicates better performance of the proposed algorithm over a competing technique.