摘要
目的肝脏肿瘤的提取是肝脏三维可视化、手术规划和模拟的基础,而当前肿瘤分割存在干预过多和分割效果不佳的问题。方法本文通过对腹部CT图像进行高斯平滑以去除图像噪声和细密纹理,计算出图像的形态学梯度并用高、低帽变换进行增强,再根据用户选择点计算内部和外部标记符,然后基于控制标记符的分水岭算法分割图像,提取出腹部CT图像中的病变组织。结果实验结果表明,该算法能够在较少的人工干预下快速分割出肝脏病变组织。结论该算法实现了腹部CT图像中肝脏病变组织的提取。
Objective Segmentation of the liver tumor regions based on abdominal CT scan images is a crucial step in 3D visualization,surgical planning and simulation. In order to solve the disadvantage of CT image segmentation and reduce the manual interaction of the liver tumor segmentation, a novel segmentation algorithm for liver tumors is proposed. Methods Firstly, the CT images are smoothed by Gaussian smoothing filter for removing noises and textures. Secondly, the gradients of the images are measured by mathematical morphology and enhanced by Top-hat transform and Bot-hat transform. Thirdly, internal marker-controlled and external marker-controlled are calculated by the points chosen by the users. Finally, the regions of liver tumors from CT scan images are extracted by the watershed algorithm based on the marker-controlled. Results Experimental results show that this method can effectively and quickly segment the liver tumors with less user interaction than other algorithms. Conclusions This method realizes the segmentation of liver tumors for abdominal CT scan images.
出处
《北京生物医学工程》
2012年第3期237-240,共4页
Beijing Biomedical Engineering
基金
湖南省教育厅优秀青年项目(09B097)
郴州市科技局计划项目(2009g106)
广东省自然科学基金团队项目(6200171)资助
