Meeting Banner
Abstract #0851

Three-Dimensional Pulmonary 1 H MRI Multi-Region Segmentation Using Convex Optimization

Fumin Guo 1,2 , Sarah Svenningsen 1,3 , Aaron Fenster 1,2 , and Grace Parraga 1,2

1 Imaging Research Laboratories, Robarts Research Institute, London, Ontario, Canada, 2 Graduate Program in Biomedical Engineering, The University of Western Ontario, London, Ontario, Canada, 3 Department of Medical Biophysics, The University of Western Ontario, London, Ontario, Canada

Many applications of pulmonary 1 H MRI require lung cavity segmentation as a prerequisite. Accordingly, we proposed a convex optimization based approach to simultaneously segment the right and left lungs from pulmonary 1 H MRI in 3D. Our approach employs the latest developments in convex optimization techniques and solves the original challenging optimization problem globally and exactly under the primal and dual perspective. We implemented the algorithm in a modern parallel computing platform and applied it to a clinical dataset of ten COPD subjects. Our experimental results demonstrate that this computationally efficient method yields highly accurate lung volumes with minimal user interaction.

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

Join Here