摘要
针对注有造影剂的双源CT心脏图像,由于血液流动和造影剂分布不均匀等使得心脏腔体存在灰度不均匀、边界泄漏等问题,采取单独的图像分割算法很难取得满意的结果,提出一种基于形态重构和Otsu多阈值的混合方法。算法首先利用形态重构运算对图像进行平滑滤波,使得心脏腔体灰度信息收敛并与周围组织产生灰度差异,再用Otsu多阈值分割算法对图像各个区域进行标记,通过指定第一张图像中注有造影剂的心脏腔体所在的标记,根据同一组织在相邻图像的关系依次识别出下一张图像中注有造影剂的心脏腔体标记,最终实现了一组心脏图像中造影剂区域的半自动分割。实验结果显示,目标区域轮廓清晰准确,为下一步心脏各腔室的分割奠定了良好的基础。
For dual-source CT （DSCT） cardiac images injected with contrast agent, due to blood flow and une- ven distribution of contrast agent, the gray-scales of the heart cavity are within proximity to each other and the gray gradients in the boundary are not obvious. Therefore, it is is difficult for single image segmentation algorithm to obtain satisfactory results. We proposed a hybrid method based on morphological reconstruction and Otsu muhi-threshold. Firstly, we used morphological reconstruction operation on image smoothing filtering, which makes the heart cavity gray-level information convergence and gray level differences with the surrounding tissue. Then the image was seper- ated into several components by Otsu multi-threshold. Finally, we picked out the regon to be segmented from first image, all of similar regions from other images can be segmented automatically according to the same group in the re- lation between adjacent images. The experimental result shows that the target area outline is clear and accurate, which paves a road for segmenting the chambers.
出处
《计算机仿真》
CSCD
北大核心
2014年第3期256-260,共5页
Computer Simulation
关键词
形态重构
多阈值
造影剂区域
Morphological reconstruction
Multi-threshold
Contrast medium area
