Brain tumour is one of the most dangerous disease occurring commonly among human beings. The chances of survival can be increased if the tumour is detected correctly at its early stage. MRI brain imaging technique is widely used to visualize the anatomy and structure of the brain. The images produced by MRI are high in tissue contrast and have fewer artifacts. It has several advantages over other imaging techniques, providing high contrast between soft tissues. However, the amount of data is far too much for manual analysis, which has been one of the biggest obstacles in the effective use of MRI. The detection of tumour requires several processes on MRI images which includes image preprocessing, feature extraction, image enhancement and classification. The final classification process concludes that a person is diseased or not. Although numerous efforts and promising results are obtained in medical imaging area, reproducible segmentation and classification of abnormalities are still a challenging task because of the different shapes, locations and image intensities of different types of tumours. In this paper, various approaches of MRI brain image segmentation algorithms are reviewed and their advantages, disadvantages are discussed.