An Efficacious Graphical User Interface Implementation for Automatic Classification of Brain Tumor from Magnetic Resonance Imaging Images Using Image Processing
Biomedical image processing is an emerging field, and is very useful for automatic disease diagnosis. Brain tumor classification is performed using image processing toolbox of MATLAB over here. For tumor classification: Image Preprocessing, Segmentation and feature Analysis are essential steps, these steps are described in the paper. First of all, images were filtered by Gaussian filter, then tumor part i.e. regions having high intensity are enhanced using power low transform in order to neglect insufficient data from image. Then enhanced part was extracted and denoising was performed to get tumor region only at output. By the method we have got only tumor part having high intensity and all other part of brain considered as background. After tumor segmentation, feature analysis has been made by which we can get idea that how they differs into normal and benign or malignant tumors. Then classified based on feature analysis of tumor region, all the processes have been done over dataset of 45 Brain Tumor images. Here overall Accuracy of correct classification is got to be 97.78%, where normal brain is classified with 100%, benign with 95% and malignant with 100% accuracy. A Graphical User Interface (GUI) is developed in MATLAB to perform all the task and to test proposed work. In future, everything would be going to be computer assisted. This is a step toward automatic disease diagnosis using Computer Aided Design (CAD).