International Journal of Biomedical Imaging 
Volume 2006 (2006), Article ID 73430, 9 pages
doi:10.1155/IJBI/2006/73430

Nonrigid Medical Image Registration by Finite-Element Deformable Sheet-Curve Models

Jianhua Xuan,1 Yue Wang,2 Matthew T. Freedman,3 Tulay Adali,4 and Peter Shields3

 1Department of Electrical Engineering and Computer Science, The Catholic University of America, Washington 20064, DC, USA
2Department of Electrical and Computer Engineering, Virginia Polytechnic Institute and State University, Arlington 22203, VA, USA
3Department of Oncology and the Lombardi Comprehensive Cancer Center, Georgetown University, Washington 20057, DC, USA
4Department of Computer Science and Electrical Engineering, University of Maryland, Baltimore County, Baltimore 21250, MD, USA
Received 16 August 2005; Revised 5 May 2006; Accepted 18 May 2006

Recommended by Ming Jiang

Abstract

Image-based change quantitation has been recognized as a promising tool for accurate assessment of tumor's early response to chemoprevention in cancer research. For example, various changes on breast density and vascularity in glandular tissue are the indicators of early response to treatment. Accurate extraction of glandular tissue from pre- and postcontrast magnetic resonance (MR) images requires a nonrigid registration of sequential MR images embedded with local deformations. This paper reports a newly developed registration method that aligns MR breast images using finite-element deformable sheet-curve models. Specifically, deformable curves are constructed to match the boundaries dynamically, while a deformable sheet of thin-plate splines is designed to model complex local deformations. The experimental results on both digital phantoms and real MR breast images using the new method have been compared to point-based thin-plate-spline (TPS) approach, and have demonstrated a significant and robust improvement in both boundary alignment and local deformation recovery.