摘要
提出了一种新的断层图像间复杂轮廓的三维重建方法。该方法基于空间距离、几何相似性和重建三维表面平滑性的准则,较好地解决了一对一和一对多分叉轮廓的匹配问题。在单轮廓匹配过程中,结合几何距离最近和双向最近邻相似准则,避免了其它算法中对阈值敏感的问题,提高了匹配的准确性和自动化程度。在一对多的分叉轮廓匹配中,区分不同的情况以进行不同的处理,得到了较满意的三维重建结果。
A novel method for 3 D reconstruction from complicated contours between slices was proposed in this paper. The method can properly solve the one-to-one and one-to-many branching contour matching problem, based on the criteria of space distance, geometrical similarity and smoothly looking of the reconstructed 3D surface. In the one-to-one contour matching process, by combining the nearest distance and bidirectional neighborhood geometrical similarity criteria, the threshold-sensitive problem that exists in other algorithms was avoided, and thus more accurate and automatic matching was achieved. For the branching contour matching, different techniques were employed for different situations, and satisfactory result was obtained in the final reconstruction.
出处
《中国生物医学工程学报》
CAS
CSCD
北大核心
2007年第3期379-383,共5页
Chinese Journal of Biomedical Engineering
基金
江苏省自然科学基金(BK2005147)
教育部科学技术研究重点项目(205060)
中科院自动化所模式识别国家重点实验室开放基金(2004)
南京大学计算机软件新技术国家重点实验室开放基金(2005)。
关键词
三维重建
轮廓匹配
三角剖分
3D reconstruction
contour matching
triangulation
