Abstract:
Brain tumor is the major cause of cancer deaths in human which is due to uncontrollable cells growth in brain portion. It is evident that the chances of survival can be increased if the tumor is detected and classified correctly at its early stage. Conventional methods involve invasive techniques such as biopsy, lumbar puncture and spinal tap method, to detect and classify brain tumors into benign (non-cancerous) and malignant (cancerous). A computer aided diagnosis algorithm has been designed so as to increase the accuracy of brain tumor detection and classification, and thereby replace conventional invasive and time consuming techniques. This study introduces an efficient method of brain tumor detection and classification, where, the real Magnetic Resonance (MR) images are classified into normal, non-cancerous (benign) brain tumor and cancerous (malignant) brain tumor. MATLAB have been used through every procedures made. These include image processing and ANN procedures. In image processing procedures, process such as image Pre-processing, histogram equalization, image filtering, segmentation, and feature extraction have been discussed in detail, followed by the methods used for classification process using ANN. Image preprocessing have been used to improve the signal-to-noise ratio and to eliminate the effect of unwanted noise. It is important to distinguish the ROI from its surroundings. This can be done by using different segmentation methods and morphological operations. The segmented ROI was considered for texture analysis. By considering the entire segmented tumor region a set of textural descriptors was calculated for each ROI using Gray Level Co-occurrence Matrices (GLCM) based second order statistics. The discriminant features that are suitable for properly differentiating the two tumor types were selected from these descriptors. The results of Co-occurrence matrices are then fed into four neural networks for further classification and tumor detection.The system was able to achieve an accuracy of 99.0%, sensitivity98.7%, specificity 100%, and an overall accuracy of classification 99.3% and of detection 99.2 %.The created systems have therefore, proved to effectively enhance the quality of the brain images and discriminate between normal and abnormal with an effective level of precision.
