Brain Tumor Segmentation from Magnetic Resonance Images Using Self-Organizing Map and Neural Network

Brain tumor is a deathly disease if not detected on time. A lot of research has been done to explore various techniques of brain tumor detection. So there are number of methods for brain tumor detection and segmentation. The timely and accurate detection and segmentation of brain tumor is very important task of any method for the convenience of doctor so that he may take necessary actions to liberate the patient’s life. But the detection and segmentation of brain tumor is very challenging work to perform due to various reasons such as blur MR images, various environmental factors, ineffective skull stripping and also limited information. But irrespective of these problems, a lot of work has been done in the field of brain tumor detection and segmentation. In this paper, we present a method of brain tumor detection from MRI images. Segmentation is done by using Self-Organizing Map (SOM) and Neural network (NN). Stationary Wavelet Transform (SWT) is used to extract the features from an input image before the training process for segmentation. We proposed a new skull stripping algorithm for the purpose of effective skull stripping. We used BRAINIX medical images as a dataset for our method. The proposed method performs better than the methods discussed in the literature. It is easy to implement and robust.