Rib Detection in 3D MRI Using Dynamic Programming Based on Vesselness and Ridgeness
In this paper, a fully automatic method is proposed to detect the ribs in 3D MRI. The purpose of the detection is MR-guided HIFU treatment of liver lesions, in which the ribs should be avoided. Rib segmentations are required for treatment planning and they may also be used for motion tracking during treatment. The rib detection results can serve as an initialization to automatic rib cage segmentation. The algorithm is based on surface detection and dynamic programming. First, the outer surface of the rib cage is detected. Vesselness and ridgeness are computed to highlight elongated structures. The ribs are tracked simultaneously on a 2D projection of the vesselness in the surface, using dynamic programming. Finally, the extracted lines are backprojected into the original 3D volume. Preliminary results of this algorithm are presented on data of five subjects. The results were evaluated by visual inspection of the backprojected lines in 3D. It was checked whether a line belonged to the correct rib and whether it stayed inside this rib. Overall, our algorithm was capable of detecting the ribs that were visible in the images. Testing on five volunteers yielded one failure. The remaining four results were satisfactory. Our method seems suitable to serve as initialization to a full rib cage segmentation in MRI.
KeywordsRibs segmentation detection MRI vesselness ridgeness dynamic programming image-guided therapy
Unable to display preview. Download preview PDF.
- 2.Klinder, T., Lorenz, C., von Berg, J., Dries, S.P.M., Bülow, T., Ostermann, J.: Automated Model-Based Rib Cage Segmentation and Labeling in CT Images. In: Ayache, N., Ourselin, S., Maeder, A. (eds.) MICCAI 2007, Part II. LNCS, vol. 4792, pp. 195–202. Springer, Heidelberg (2007)CrossRefGoogle Scholar
- 6.Lorigo, L.M., Faugeras, O., Grimson, W.E.L., Keriven, R., Kikinis, R.: Segmentation of Bone in Clinical Knee MRI Using Texture-Based Geodesic Active Contours. In: Wells, W.M., Colchester, A., Delp, S. (eds.) MICCAI 1998. LNCS, vol. 1496, pp. 1195–1204. Springer, Heidelberg (1998)CrossRefGoogle Scholar
- 8.Zhang, K., Lu, W.: Automatic Human Knee Cartilage Segmentation from Multi-contrast MR Images Using Extreme Learning Machines and Discriminative Random Fields. In: Suzuki, K., Wang, F., Shen, D., Yan, P. (eds.) MLMI 2011. LNCS, vol. 7009, pp. 335–343. Springer, Heidelberg (2011)CrossRefGoogle Scholar