Tensor Field Regularization Using Normalized Convolution

Abstract. This paper presents a filtering technique for regularizing tensor fields. We use a nonlinear filtering technique termed normalized convolution [Knutsson and Westin 1993], a general method for filtering missing and uncertain data. In the present work we extend the signal certainty function to depend on locally derived certainty information in addition to the a priory voxel certainty. This results in reduced blurring between regions of different signal characteristics, and increased robustness to outliers. A driving application for this work has been filtering of data from Diffusion Tensor MRI.

• Type: Conference
• Year: 2003
• Where: EUROCAST
• Authors: Carl-Fredrik Westin, Hans Knutsson
• Keywords: EUROCAST 2003, Nonlinear filtering Technique, Priory Voxel Certainty, Signal Certainty Function