Comparative study of Image Segmentation methods in detection of Brain Tumour

Image segmentation of MRI images is done to detect brain tumour. Image segmentation is the partition of the image into a number of segments which enables us to study the image in a meaningful manner. It gives us advanced insight of the image which without segmentation would be extremely difficult. This paper presents various methods of detecting these tumours. We will be detecting the tumour using three methods: Threshold method, Watershed method and K-Means method. We use a training mechanism which consists of multiple MRI images of the brain with the tumour which helps in accurate detection of brain tumour. Multiple parameters like Mean, Variance, Homogeneity, Skewness, correlation, etc are determined. These parameters in turn help us to classify the tumour into Malignant or Benign. Malignant tumour is cancerous and Benign is non-cancerous. In this paper, we provide a means to compare the three mentioned methods of image segmentation and to understand which one is best suited for the application. To carr this out, we use MATLAB for running the algorithm for all the three methods and use GUIDE, to create a Graphical User Interface in MATLAB to display the result of the brain tumour detected