PERFORMANCE OF MRI BRAIN TUMOR CLASSIFICATION USING NEURAL NETWORK CLASSIFIERS

In Medical Image processing Brain tumer classification plays vital role to diagnosis system for Magnetic Resonance Imaging analysis. As traditional approaches performs low classification accuracy due to its inadequate functionalities to handle the medical image datasets having multiple classes. In this paper the work presents an effective algorithm to classify a tumor in brain MRI images using intensity-based statistical features and deep neural network. Data, within the region of interest (ROI), features extraction based on gray-level co-occurrence matrix (GLCM) and BT segmentation based on Discrete Wavelet Transform (DWT). These filters are combined in this algorithm as directional transformation methods for exploiting all information in all directions of the MRI input image. MRI Features are extracted based on the first and second order statistics from both domains. Furthermore this work incorporates two types of neural network classifiers: Stacked Sparse Auto encoder (SSA) and Softmax Classifier (SMC). BRATS dataset consist of 6643 MRI images and classified into three tumor classes such as: Pituitary, Glioma, and Meningioma. We have evaluated the classification accuracy with traditional classifiers such as support vector machine, K Nearest Neighbor Classifier and Logistic regression: The performance of the proposed algorithm is validated using the experimental results in terms of accuracy, specificity, and sensitivity compared to the existing algorithms. The highest accuracy obtained is 98.8 % by the proposed classifier on BRATS dataset respectively.