Automatic Brain Tumor Identification using Clustering of K-means Algorithm in Image Processing

M Praveena, G Rohini, G Tejeswara Reddy, K Hemanth Sai Nikhil

In this paper, one of the simplest detection algorithms called clustering of k-means is used to identify the brain tumor. The brain tumor is a clot-shaped disease that attacks the brain which is difficult to distinguish brain tumor tissue from normal tissues such as fat and cerebrospinal fluid due to similar color. CT scans are commonly used to identify the brain tumors. CT scans use radiation to create an image of the brain. Because of this, it affects the human body and it is time taking process because it takes 24 hours to generate the report from the extracted images during the examination in CT scanner. To overcome the difficulty, MRI images are used to overcome this problem so that the human body is not affected because it uses radio signals and electromagnetic waves. The MRI picture is first added with noise by using “salt and pepper” technique and then preprocessed using a Gaussian filter technique which eliminates the noise in the image and produce smoothness to the image. The final method is segmentation method and it is performed to identify the region of the tumor and its shape using kmeans clustering. Since the tumor region has similar pixels, the area is created as a cluster by clustering. The clustered area is identified as white in color and remaining is black in color, the white colored area is the detected tumor. The segmentation results are a contrast between area of brain tumor and area of brain tissues.

Volume 11 | Issue 7

Pages: 621-630