提出一种适合于连续医学图像序列的轮廓匹配算法。基于笛卡尔坐标空间,首先对医学图像轮廓曲线进行离散化,运用多边形逼近算法提取点数最多轮廓的特征点,然后依据多边形相似度来提取剩余轮廓的特征多边形,使得相邻轮廓之间达到最好的匹配。实验结果表明,该算法可以得到很好的匹配效果。
This paper presents a matching method fitting for many contours group of successive medical image sequences. The algorithm is based on the Descartes coordinate space. Firstly, extract the feature polygon of discrete curve, and then calculate the maximal similarity between adjacent feature polygons, finally satisfy the principle of similarity. The experimental results show that the algorithm can get very good match results.