A Novel Classification of MRI Brain Images using ANFIS and 3D Reconstruction
M. Fathima Zahira1, M. Mohamed Sathik2

1M. Fathima Zahira, Research Scholar, Bharathiar University, Coimbatore, Tamil Nadu, India.
2Dr. M. Mohamed Sathik, Principal, Sadakathullah Appa College, Tirunelveli, Tamil Nadu, India.
Manuscript received on September 23, 2019. | Revised Manuscript received on October 15, 2019. | Manuscript published on October 30, 2019. | PP: 5434-5440 | Volume-9 Issue-1, October 2019 | Retrieval Number: A3082109119/2019©BEIESP | DOI: 10.35940/ijeat.A3082.109119
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: The brain tumor (BT) has turn into a chief hazard to life in many humans. With the advances in medicine and technology, early tumor detection may pay a way for treating it in an early phase and thereby reducing the death rates. MRI imaging has a significant role in imaging the BT. Neurologist base the treatment of BT on the type, location, and also size of the tumor and hence proper segmentation of the tumor region has become essential. Here, an efficient segmentation algorithm is proposed, is centered on using the Sobel edge detection mechanism and the classification of the tumor region centered on the features extorted as of the segmented images are performed. The brain MRI images from a database which contain normal and also abnormal cases. These images are stripped from the skull by utilizing the morphological operations like morphological opening along with closing. Following this, segmentation is executed and finally, features are extorted as well presented to the classifier where the classification is executed. The segmentations algorithm is estimated and also the outcome are contrasted to the other prevailing algorithms say Kmeans and SVM (Support Vectors Machine) and the classification algorithm is weighted against that of the existent classification algorithm like PNN (probabilistic neural network) and ANN (Artificial Neural Network). Thus the proposed algorithms are confirmed to be superior to the other algorithms used in BT segmentations and classification.
Keywords: Sobel edge detection, Support Vector Machine, Kmeans algorithm, morphological operation, probabilistic neural network, Artificial Neural Network.