Medical Image Denoising by Nonlocal Means with Level Set Based Fuzzy Segmentation

Objectives: Brain tumor segmentation is an important challenge in the medical image processing. There is a need for early diagnosis of brain tumor. It enhances the treatment possibilities based on their diagnosis results. The manual segmentation of brain tumor is a challenging task because large numbers of MRI images are generated and is difficult to analyse by the radiologist and also consuming more time. To overcome this problem an automatic brain tumor image segmentation is developed. Methods/Statistical Analysis: This comprises of pre-processing and segmentation of MRI images. First, Enhanced Nonlocal Means Filter (ENLMF) is proposed to remove the noise from the brain images. The ENLMF algorithm is developed to prevent the heavy computation load generated by the standard NLMF algorithm and is a popular method for denoising the medical images. Second, fuzzy level set algorithm is proposed to segment the medical images. The initial segmentation is obtained through fuzzy clustering and controlling parameters are generated from the levels set algorithm and is useful for robust segmentation purpose. Findings:The performance of the proposed system is evaluated using metrics such as PSNR, MSE, RMSE and quality metrics and the images are collected from online websites.