Analysis and evaluation of classification and segmentation of brain tumour images

摘要

Apparently, the development of a model to detect the tumour part in brain images is of utmost significance. In the initial phase of our work, brain tumour database images have occurred in the preprocessing module using adaptive median filter technique to gain clarity of the image. In addition to the preprocessing process, feature extraction techniques are applied and extracted to the features then the classification method as support vector machine (SVM) classifier is used to classify the images as normal and abnormal. After classification, the abnormal images are observed for segmentation process using fuzzy c-means (FCM) clustering process along with the occupied optimisation methods. For optimising centroid, the FCM used social spider optimisation (SSO) technique with genetic algorithm. The proposed scheme has attained the maximum accuracy when compared to existing classification technique ANFIS and segmentation technique FCM (GWO).