Abstract: Three-dimensional reconstruction technology is an effective tool which can help doctors analyze and show the state of an illness and it greatly improves the accuracy of medical diagnosis.The marching cubes is a classical algorithm for three-dimensional reconstruction of medical images，but it still shows some disadvantages.Targeting to two disadvantages which are structuring surfaces by line instead of hyperbola and spending much time detecting null cubes in the traditional algorithm，this paper implements an improved MC algorithm and solves two problems by interpolating between two images and searching based on image segmentation.Comparing the traditional algorithm，experiments show that the improved MC algorithm achieves great enhancement in effect and efficiency.

摘要： CT三维重建技术是辅助医生对病情进行分析和显示的有效工具，它极大地提高了医疗诊断的准确性。移动立方体法（Marching Cubes，MC）是一种经典的医学图像三维重建算法，但是在实际使用中还是存在着一些缺点。针对传统MC算法出现的两个问题：（1）用直线代替双曲线来构造等值面会产生误差；（2）重建过程中大量时间耗费在检查空立方体上，提出了一种改进的MC算法，分别通过切片图像间的插值和基于分割的快速查找法来对此进行解决。通过实验表明，改进的MC算法比起传统MC算法，在三维重建的效果和效率上都得到了很大的提高。