Multimodal Magnetic Resonance Image Restoration Using a Vector-Valued Anisotropic Diffusion-Based Mixed Noise Filtering Technique

A medical multimodal magnetic resonance image (MRI) restoration solution that removes successfully the mixtures composed of Gaussian and quantum noises is introduced in this article. The proposed approach is based on a nonlinear vector- valued second-order anisotropic diffusion model. This model determines a nonlinear reaction- diffusion equation system whose single-valued diffusion equations share some correlation terms. A consistent finite difference-based numerical algorithm that solves this system is proposed next. It is applied successfully in the 3D MRI filtering experiments that are finally described here.