A Novel Approach to Find Tumor in MRI Brain Images using Image Segmentation Techniques

Segmentation is one of the most important task in image processing. It classifi es the pixels into two or more groups depend on their intensity value. The quality of the segmentation is based on the method applied to select the algorithm. The objective of the research work is to fi nding the tumor affected region by extending and modifying the Electromagnetic optimization algorithm with the level set function combined with the active contours. The new hybrid ELSM algorithm is composed of three stages: preprocessing stage and the second stage is evaluating the modifi ed EMO algorithm for segmenting the images and the last stage is to implementing the boundary detection technique to enhance the tumor affected region separately. The fi rst stage is converting DICOM images into standard image fi le formats and preprocessing with the liner fi lter to remove noise and adjusting contrast levels of images. The second stage is implementing modifi ed EMO algorithm for segmenting the images with the attraction and repulsion mechanism for the intensity value of the images with the cluster value of n and the last stage is detecting boundary region of the segmented images and its been evaluated with the modifi ed level set functions and with the of level values k to obtain the best optimal images for the MR brain tumor problems. The images are collected and the ELSM algorithm been evaluated and the results are discussed. The ELSM algorithm best resultant images are obtained by analyzing the n and k values of the new hybrid algorithm using maximum and minimum techniques.