An Effective Brain Tumor Detection Using Modal Based Classification

Abstract

Brain tumor is nothing but growth of abnormal cells in the brain. In the initial stages of brain tumor indications like heavy headaches, weakening of vision occur. Sometimes there were no indications also. So, adequate processing of the image helps to detect the tumors at the initial stages. In medical operations, image processing is one of the most demanding techniques. Hence in this project we mainly focus on the preprocessing of image, segmentation of image and the classification of features in the MRI images of the brain. Till today we don’t know how to remove the brain tumor. But if we identify it in the initial stages, we have more chances to handle it. Survival of the affected person is also high. But this task is time taking and can be performed by clinical experts and the radiologists only. The accuracy of the detected tumor is totally depending on the experience of the experts. The main functions like enhancement, segmentation and classification can be done by computer-based technology using model-based classification. It is done on image data. These are used in wide range of real-time applications. These consist of computer oriented photos, model-based classification and machine learning related to therapeutic, biological and remoteness. Analysis of dis- ease can be done by the surgeon using this model. To increase the MRI image quality, we use image enhancement and then we will pre-process the MRI image. To detect tumor tissues from MRI images segmentation is used. Segmentation process is the separation of the image into blocks which have the same properties like texture, boundaries, color, gray level and brightness. This process involves separation of normal tissues from abnormal tissues in the image. Morphological operations performed on the segmented part. Finally, to compare the features to testing data with training data we use classifiers