Motion Correction for Dynamic Contrast-Enhanced CMR Perfusion Images Using a Consecutive Finite Element Model Warping
We present results of a non-rigid registration algorithm to correct breathing motion in cardiac MR perfusion sequences applied to the STACOM 2014 Motion Correction Challenge dataset. The algorithm is based on the finite element method whereby a 2D free form deformation model is deformed to match image features. Image warping is performed through global-to-local mapping of motion parameters. To overcome the contrast intensity problem in the perfusion images, the registration was applied consecutively between adjacent frames. Eleven cases were provided by the challenge: Ten cases were ECG-gated MR perfusion images with rest and adenosine-induced stress series, while the last case was an ungated MR perfusion stress acquisition. The algorithm achieved good results in terms of modified Hausdorff distance: \(1.31\pm 0.93\) pixels (max: 6.5 pixel), horizontal shifting \(< 4.5\) pixels, and vertical shifting: \(< 4\) pixels. Moderate Jaccard index: \(0.57\pm 0.14\) was achieved.
Unable to display preview. Download preview PDF.
- 3.Dubuisson, M.P., Jain, A.: A modified Hausdorff distance for object matching. In: Proceedings 12th IAPR International Conference on Pattern Recognition - Conference A: Computer Vision & Image Processing, vol. 1, pp. 566–568 (1994)Google Scholar
- 6.Li, B., Young, A.A., Cowan, B.R.: GPU accelerated non-rigid registration for the evaluation of cardiac function. Med. Image Comput. Comput. Assist. Interv. 11(Pt 2), 880–887 (2008)Google Scholar
- 7.Press, W.H.: Numerical recipes in C: the art of scientific computing, 2nd edn. Cambridge University Press, Cambridge (1992)Google Scholar