Image Segmentation and Classification of MRI Brain Tumor Based on Cellular Automata and Neural Networks

This paper proposes segmentation of MRI brain tumor using cellular automata and classification of tumors using Gray level Co-occurrence matrix features and artificial neural network. In this technique, cellular automata (CA) based seeded tumor segmentation method on magnetic resonance (MR) images, which uses volume of interest (VOI) and seed selection. Seed based segmentation is performed in the image for detecting the tumor region and then highlighting the region with help of level set method. The brain images are classified into three stages that are normal, benign and malignant. For this non knowledge based automatic image classification, image texture features and Artificial Neural Network are employed. The conventional method for medical resonance brain images classification and tumors detection is by human inspection. Decision making was performed in two stages: feature extraction using Gray level Co-occurrence matrix and the classification using Radial basis function which is the type of ANN. The performance of the ANN classifier was evaluated in terms of training performance and classification accuracies. Artificial Neural Network gives fast and accurate classification than other neural networks and it is a promising tool for classification of the tumors.