摘要
图割理论作为一种全局最优的框架,为腹腔CT图像肝脏提供了一种可行化方法。利用图割理论和先验知识进行CT图像肝脏分割首先需要从CT图像中得到肝脏感兴趣区域,再利用CT图像中存在的关于肝脏的先验知识辅助地选取种子点,减少了交互时间,最后运用最大流最小割算法实现肝脏分割并经过后处理得到较准确的肝区。实验结果表明该方法分割结果准确,执行效率高,鲁棒性好。
As a global optimum theory, the Graph Cut method provides a framework for liver segmentation in abdominal CT im- ages. In this paper, a method based on Graph Cut is proposed to segment liver from CT image with the usage of prior knowl- edge. Firstly, R.OI （Region of Interest） of the CT image is obtained to lower the memory footprint and the computation time; Secondly, prior knowledge in CT image is applied for seeds selection to decrease the interactive time; Finally, the max-flow/ rain-cut algorithm is utilized to segment the liver, subsequent post-processing steps further refine the segmentation. Experiments demonstrate the robustness of the proposed method with good accuracy and high performance.
出处
《电脑知识与技术》
2012年第12期8300-8303,8310,共5页
Computer Knowledge and Technology
关键词
肝脏CT图像
图割
分割
先验知识
鲁棒性
Liver CT images
Graph Cut
segmentation
prior knowledge
robustness
