Primitive prognosis of brain disease in medical images using multi-model textures features and fuzzy logic oriented kernel SVM

Abstract

Image segmentation plays a primitive and indispensable step in medical image processing. The MR imaging has
become a broad spectrum of utilizing high quality medical imaging, especially in brain imaging where the soft-tissue
contrast and non-invasiveness is a clear advantage. In this paper a novel TCM (Texon Co-occurrence matrix) based brain
tumor classification system has been designed and developed for MRI systems. The proposed method incorporates three
stages namely pre-processing, feature extraction and classification. In pre-processing Anisotropic filter is applied for
removing the noise for experimental image, In the second stage histogram and Texon Co-occurrence matrix are used for the
purpose of feature extraction .At Final step the supervised Fuzzy SVM (Support Vector Machine) classifier is used to
classify the type of tumor image. The extracted features are compared with the stored features in the knowledge base and
the type of tumour is classified. Thus ,the proposed system has been evaluated using the metrics sensitivity ,specificity and
accuracy .The proposed system was found efficient in classification with a success of more than 95 % of accuracy.