Lab. Traitement du Signal et de l'Image, Rennes I Univ.
IEEE Trans Med Imaging. 1991;10(2):122-31. doi: 10.1109/42.79469.
An approach to the three-dimensional reconstruction of coronary arteries is presented. The principal objective is to show how modeling of a vascular network, together with algorithmic procedures, can lead to accurate 3-D structure and feature labeling. The labeling problem is stated directly within the 3-D reconstruction framework. The reconstruction ambiguities inherent to biplane techniques are solved by means of a knowledge base, modeling of the object, and heuristic rules. Feasibility in near-real situations has been demonstrated. The critical importance of the object 3-D reference to achieving the data and modeling matching is emphasized, and a way to deal with it is pointed out. The overall system implies an incremental development in methodologies and experiments. All of them have been elaborated and tested independently, and the most appropriate ones have been selected for integration into a modular system. All the stages of the process (calibration, segmentation, reconstruction, and display) are discussed, with the main focus on modeling. Examples of automatic reconstruction from a phantom are provided.
提出了一种冠状动脉的三维重建方法。主要目的是展示如何通过血管网络建模以及算法过程来实现准确的三维结构和特征标记。标记问题直接在三维重建框架中提出。通过知识库、对象建模和启发式规则来解决双平面技术固有的重建歧义。已经在近乎真实的情况下证明了其可行性。强调了对象三维参考对于实现数据和模型匹配的重要性，并指出了一种处理方法。整个系统意味着在方法和实验方面的逐步发展。所有这些都已经独立地进行了阐述和测试，并且已经选择了最合适的方法来集成到一个模块化系统中。讨论了该过程的所有阶段（校准、分割、重建和显示），主要重点是建模。提供了从体模自动重建的示例。
