用于约束非刚性配准算法的器官重叠跟踪
TRACKING ORGAN OVERLAP FOR A CONSTRAINED NON-RIGID REGISTRATION ALGORITHM.
作者信息
Greene W H, Chelikani S, Papademetris X, Staib L H, Knisely J P S, Duncan J
机构信息
Department of Biomedical Engineering, Yale University, New Haven, CT, USA.
出版信息
Proc IEEE Int Symp Biomed Imaging. 2008;4541207:1159. doi: 10.1109/ISBI.2008.4541207.
This paper tracks organ (prostate, rectum, bladder) overlap in a constrained non-rigid registration (NRR) algorithm to register computed tomographic (CT) images used in external beam prostate radiotherapy. The local motion of the organs is described by a hierarchical multi-resolution FFD based on cubic B-splines. Registration is achieved by minimizing a cost function which is a combination of three functions representing the overlap of the critical organs, image similarity and smoothness of the transformation. The constrained NRR algorithm generated better registration results when compared to an unconstrained NRR algorithm.
本文在一种用于配准适形调强前列腺放疗中使用的计算机断层扫描（CT）图像的约束非刚性配准（NRR）算法中追踪器官（前列腺、直肠、膀胱）重叠情况。基于三次B样条的分层多分辨率自由变形（FFD）描述了器官的局部运动。通过最小化一个代价函数来实现配准，该代价函数是代表关键器官重叠、图像相似性和变换平滑度的三个函数的组合。与无约束NRR算法相比，约束NRR算法产生了更好的配准结果。
Zotero 插件