Abstract #3573

Denoising Diffusion-Weighted MR Images Using Low Rank Structure and Edge Constraints

Fan Lam¹,², S. Derin Babacan², Norbert Schuff³, Zhi-Pei Liang¹,²

¹Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL, United States; ²Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL, United States; ³Center for Imaging of Neurodegenerative Diseases, VA Medical Center, San Francisco, CA, United States

A novel method is proposed to denoise diffusion weighted image sequences. The proposed method uses a penalized maximum likelihood formulation that handles Rician noise and incorporates low rank structure and prior edge information. The proposed method has been evaluated using experimental DTI data, and provides superior performance on recovering image features, anisotropy and orientation information of diffusion tensors originally corrupted by noise. We expect the proposed method to prove useful for achieving higher measurement precision and/or reducing data acquisition time for diffusion MRI.