Sierra R, Zsemlye G, Székely G, Bajka M
Department of Information Technology and Electrical Engineering, Computer Vision Laboratory, Sternwartstrasse 7, 8092 Zurich, ETH Zürich, Switzerland.
Med Image Anal. 2006 Apr;10(2):275-85. doi: 10.1016/j.media.2005.11.003. Epub 2006 Jan 4.
The generation of variable surgical scenes is a key element for effective training with surgery simulators. Our current research aims at a high fidelity hysteroscopy simulator which challenges the trainee with a new surgical scene in every training session. We previously reported on methods able to generate a broad range of pathologies within an existing healthy organ model. This paper presents the methods necessary to produce variable models of the healthy organ. In order to build a database of uteri, a volunteer study was conducted. The segmentation was carried out interactively, also covering the establishment of an anatomically meaningful correspondence between the individual organs. The variability of the shape parameters has been characterized by principal component analysis. A new method has been developed and tested, allowing the derivation of realistic new instances based on the stochastic model and complying with non-linear shape constraints which are defined and interactively controlled by medical experts.
生成可变的手术场景是使用手术模拟器进行有效训练的关键要素。我们目前的研究旨在开发一种高保真宫腔镜模拟器，在每次训练中都为学员带来全新的手术场景挑战。我们之前报道过能够在现有的健康器官模型中生成多种病变的方法。本文介绍了生成健康器官可变模型所需的方法。为了建立子宫数据库，我们开展了一项志愿者研究。分割是交互式进行的，还包括在各个器官之间建立具有解剖学意义的对应关系。形状参数的变异性已通过主成分分析进行了表征。我们开发并测试了一种新方法，该方法允许基于随机模型推导逼真的新实例，并符合由医学专家定义和交互式控制的非线性形状约束。
