Classification and Validation of MRI Brain Tumor Using Optimised Machine Learning Approach

摘要

The influence and impact of digital images on modern society, science, technology and art are incredible and image processing is now a critical part of science and technology. A brain tumor is a distinctive and abandoned enlargement of brain cells, which are the source of death associated with cancer. Early detection of the brain tumors will cut the unconditional deaths of young people. Detection of brain tumor is complex because of the complex size of the brain. MRI (Magnetic Resonance Images) can give detail information with respect to the tissue life systems, which is for the recognition of brain tumors. Distinctive phases are included for the recognition of Brain Tumor i.e. preprocessing, segmentation, feature extraction and classification. Diagnostic MRI system corresponds to automated system involving enhancement of segmentation and classification process is discussed in this paper. The segmentation is the initial step that segments the benign and malignant tumor by utilizing filtering techniques available in image processing and then the classification approach to be executed. Modified median filter and multi-vector segmentation machine is used to form the segmented tumor region in the images. At the last stage, the implementation of the suggested techniques evaluated with multi support vector algorithm which distinguishes the tumor and MRI images. The proposed method efficiency increased in terms of RBF accuracy and linear accuracy. The performance analysis shows 10% betterment as compared to system exclusive of the application of modified median filtering with intensity adjustment feature.