Morphology Based Brain Tumor Identification and Segmentation in MR Images

The human body comprises various types of cells where the brain plays a very important role. Due to the complex nature of brain structure, the use of the Magnetic Resonance Imaging (MRI) technique has grown extensively. MRI is used to scan and examine the anatomical structures of the human brain & produces high-resolution images that are rich in information and therefore reveal salient characteristics of the brain. These images are analyzed to extract the clinically important information required for medical diagnosis and biomedical research. In this study, a morphology-based image processing technique is proposed for the tumor segmentation in brain MR images. Firstly, the brain MRI images are preprocessed using an anisotropic diffusion filter (ADF) for eliminating high frequency noise while preserving the edges of existing objects. Secondly, filtered images go through the skull stripping process to remove the scalp tissue, skull and dura. And finally, the properties of the skull stripped image is measured to identify the presence of the tumor. In case of tumor existence, the morphology-based technique is performed on brain images to segment the tumor. It is found that the proposed system is able to identify tumorous images correctly and extract the features of the segmented tumor.