Projects:LMMSERicianDWINoiseRemoval

Back to NA-MIC_Collaborations

Objective: Provide a way for noise removal in diffusion weighted images incorporating the Rician noise model. The Rician noise level is estimated automatically and used to parametrize the local Linear Minimum Mean Squared Error estimator.

Progress: The method was evaluated on real and synthetic datasets. A Slicer 3 module was developed.

References:

• Aja-Fernandez, S., Alberola-Lopez, C., Westin, C.F., "Filtering and noise estimation in magnitude MRI and Rician distributed images," submitted to IEEE Transactions on Image Processing.
• Aja-Fernandez, S., Niethammer, M., Kubicki, M., Shenton, M.E., Westin, C.-F., "Restoration of DWI data using a Rician LMMSE estimator," submitted to MRM.

Key Investigators:

• Santiago Aja-Fernandez, Marc Niethammer, Marek Kubicki, Martha Shenton, Carl-Fredrik Westin.

Example results:

FA values and direction of major tensor eigenvalue based on original DWI data.

FA values and direction of major tensor eigenvalue after filtering the DWI data using the LMMSE filter.