Extraction and description of tumour region from the brain MRI image using segmentation techniques

Image segmentation refers to partitioning of image into multiple regions (segments) eventually leading to meaningful representation of image through which information can be extracted. Magnetic Resonance Imaging (MRI) is a non-invasive method which provides the detailed view of internal structure of tissues in the body at very high resolution. This paper focuses on extraction of brain tumor and its region description through segmentation from the brain MRI image. At first, pre-processing step for noise removal is carried out. Brain tumor extraction is done by considering the methods based on k-means clustering, morphological operations and region growing. A comparative analysis of the three methods is done for various brain MR images. Here we refer the tumor in the brain MRI as region of interest (ROI) and rest of the image is referred as Non-region of interest (NROI).