Segmentation and Area Calculation of Brain Tumor Images Using K-Means Clustering and Fuzzy C-Means Clustering

Tumors of the brain are abnormal growths of cells that are precursors to cancer and may lead to low survival rates. So, the early detection of tumors can be a life saving measure. Brain tumor can be seen using MRI image but the problem is to differentiate the tumor from the normal tissue. This can be overcome by image segmentation. To know the exact location and size of the tumor, K-means and Fuzzy C-means (FCM) clustering and Fuzzy C means plus thresholding combined segmentation techniques are used in this work. And the results of these methods are compared. First the MRI images are exposed using image preprocessing techniques like median filter, skull stripping etc. and then the preprocessed images are subjected to segmentation and feature extraction is done by the proposed methods. The performance of segmentation methods is evaluated using PSNR and IMMSE. The performance results show that Fuzzy C means with the thresholding method is giving better results compared to other segmentation methods. Finally. the area of the tumor is calculated by counting the number of pixels in the tumor region after some morphological operations.