Adaptive Spatial Shrinkage Estimation of Diffusion Tensor Imaging with DWI Data

• Activity Number: 335
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Tao Yu*+
• Companies: National University of Singapore
• Keywords: Adaptive Kernel Smoothing ; Diffusion Weighted Imaging ; DTI ; L1 Penalty ; Locally Weighted Least Squares

Abstract

Diffusion Tensor Imaging (DTI) data are typically estimated from Diffusion Weighted Imaging (DWI) data acquired from magnetic resonance experiments. We propose a two-stage adaptive spatial shrinkage estimation (ASpSkE) procedure for estimating DTI data from DWI data concerning the following two aspects. First, incorporate the spatial information in the diffusion tensor (DT) estimation by an adaptive kernel smoothing approach. Second, reduce the systematic bias components in the estimated eigenvalues of DTs typically present in existing approaches in DT estimation. The effectiveness of ASpSkE is demonstrated by simulation and real data examples.