Gradient Magnitude Based Watershed Segmentation for Brain Tumor Segmentation and Classification

摘要

MRI is one of the tool for detecting the tumor in any part of the body. But precise tumor segmentation from such Magnetic resonance imaging (MRI) is difficult and also time consuming technique. To overcome such difficulty, this work proposes a very simple, efficient and automatic segmentation and classification of brain tumor. The proposed system is composed of four stages to segment, detect and classified tumor as benign and malignant. Pre-processing is carried out in the first stage after which watershed segmentation technique is applied for segmenting the image which is the second stage. The segmented image undergo for post processing to remove the unwanted segmented image so as to obtain only the tumor image. In the last stage, gray-level co-occurrence matrix (GLCM) is used to extract the feature. This feature is given as input to Support Vector Machine (SVM) to classify the brain tumor. Results and experiment shows that the proposed method accurately segments and classified the brain tumor in MR images.