Multispectral brain Magnetic Resonance Imaging denoising using non-local means with statistical nearest neighbor

The basic operation of computer vision is filtering. It is the overture process in applications of biomedical image processing. MRI denoising improves the diagnosis accuracy, quality of image with operations performed on it such as image segmentation and registration; and also parameter estimation. It has been observed that the noise in MRI follows gaussian and rician distribution. In this paper, algorithms are applied on multispectral brain MRI that are either T1-weighted or T2-weighted. The proposed method Multispectral non-local means with statistical nearest neighbor is applied and the results are compared with other NLM filters available. The experimental results show that proposed method provides better noise reduction than NLM with nearest neighbor and adaptive NLM.