A Novel Image Segmentation Approach for Brain Tumor Detection Using Dual Clustering Approach

摘要

Detection of brain tumors is a major issue nowadays due to the diagnosis of exact location and dimension of those tumors. This paper approaches image segmentation for brain tumor detection using dual clustering technique. Here the input image is separated by three channels of Hue, Saturation and Brightness and computing threshold for each channel to create a bitmap with the help of largest threshold computed from the three channels. Then the bitmap is segmented into black and white pixels to find out the nonoverlapping connected set of pixels and spots. The process of computing threshold includes the recognition of gray areas in the saturation and brightness channels.