Analysis of brain MRI images for Tumor segmentation using Fuzzy C means Algorithm

In general, the amount of information present in biomedical images is more and it is a very challenging task to handle these kinds of images in wireless communication. Image retrieval is one of the most emerging technologies for telemedicine to handle the medical data of images (MRI and CT). Brain tumor segmentation is that the vital method to portrait the beginning level of tumor. Magnifying the tumor is being an enormous challenge due to the complex characteristics of the MRI Images which provides high intensive, divergent and uncertain boundaries. In this paper, the tumor present in the brain MRI is segmented using a new formulation technique of FCM (fuzzy C means algorithm) called PIGFCM algorithm has been introduced, followed by an image de-noising technique, which is an inevitable pre-processing step in image processing. Image de-noising technique is employed using PSNLM filter (i.e. Pre-Smooth Non-Local Means filter), which is used for denoising the Rician noise (noise present in the image, which is difficult to remove), and the fuzzy algorithm used in the proposed method is called as the PIGFCM algorithm, which is a reformulation of the FCM (fuzzy C means algorithm), which incorporates good quality in the segmentation process. PIGFCM algorithm utilizes the prior information of the tumor classes. The proposed technique has better de-noising results, substantial superior segmentation accuracy and good speed. The PSNR and the NMSE of the results are also calculated.