Meeting Banner
Abstract #0781

Discontinuity Preserving Registration using Truncated L1 Regularization and Minimum Spanning Tree based Motion Clustering

Dongxiao Li1,2, Juerong Wu1, Kofi M. Deh2, Thanh D. Nguyen2, Martin R. Prince2, Yi Wang2,3, and Pascal Spincemaille2

1College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou, China, People's Republic of, 2Department of Radiology, Weill Cornell Medical College, New York, NY, United States, 3Department of Biomedical Engineering, Cornell University, Ithaca, NY, United States

Free breathing liver perfusion analysis requires non-rigid motion registration of the unavoidable respiratory motion in the dynamic data. Traditional non-rigid methods rely on spatially smooth motion parameters, which is problematic for the sliding motion of the liver against the abdominal wall. In this work, truncated L1 regularized Minimum Spanning Tree based motion clustering combined with a Markov Random Field optimization is proposed to perform liver registration without the need for manual segmentation. Results on breath-hold liver images acquired at various positions of the respiratory cycle demonstrated this method allows superior liver motion estimation when compared to traditional methods.

This abstract and the presentation materials are available to members only; a login is required.

Join Here