作　　者： ; ; ;

机构地区： 天津医科大学生物医学工程学院

出　　处： 《计算机应用》 2011年第3期760-762,共3页

摘　　要： 为提高图割算法的效率并减少用户交互量,提出将图割与均值漂移算法结合应用的脊椎骨自动分割方法。该方法利用均值漂移算法产生的区域邻接图代替像素点图,从而大幅减少参与图割算法的顶点和边的数目,并有效利用了均值漂移良好的边界结构保持特性。实验结果表明,该方法有效地结合了两者的优点,提高了算法的精度和速度,并减少了用户交互量。 To improve the efficieney of graph eut algorithm and reduce user interaction, the graph cut and mean shift algorithm were combined to automatically segment 3D vertebra images. The image was pre-processed with the mean shift method and the region graph was obtained instead of pixel graph, which could reduce dozen times edge numbers in the graph. Furthermore, the edge-preserving property of mean shift algorithm was used. The experimental results show that the proposed method possesses nice properties of the mean shift and graph algorithm, so it is efficient and accurate, and it also reduces user interaction.

关 键 词： 图割 均值漂移 脊椎骨分割

领　　域： [自动化与计算机技术] [自动化与计算机技术]