您的位置：首页 > 论文页面
基于Snake的颈动脉内中膜厚度提取
发表时间：2014-02-28 浏览量：1819 下载量：520
| 全部作者： | 吴慧慧，吴琼，徐红卫，刘洋，王犀点，杨鑫，丁明跃 |
| 作者单位： | 华中科技大学生命科学与技术学院，图像信息处理与智能控制教育部重点实验室；郑州大学第五附属医院CT室；华中科技大学多谱信息处理技术国防科技重点实验室，图像识别与人工智能研究所 |
| 摘 要： | 提出了一种基于主动轮廓模型的自动膜提取法（automated layer extraction based on Snake，ALES）提取超声颈动脉内中膜厚度。在选定感兴趣区域（region of interest，ROI）并得到初始轮廓后，超声颈动脉内膜用Snake方法迭代，外膜用梯度矢量流（gradient vector flow, GVF）作为外力的GVF-Snake方法进行迭代，得到最终内外膜轮廓。而后，将其与一种通过使用新的外部能量函数改进的活动模型方法进行对比，从而对其进行验证。验证实验基于已知金标准数据的超声颈动脉图像，通过计算改进的活动模型方法和ALES得到的内中膜厚度（intima-media thickness，IMT）与金标准的相对误差对2种方法的分割结果进行对比。结果表明，ALES得到的IMT与金标准的相对误差更小，故在超声颈动脉内中膜厚度的提取中可得到更加准确的结果。 |
| 关 键 词： | 计算机图像处理；图像分割；基于主动轮廓模型的自动膜提取法；颈动脉内中膜厚度 |
| Title： | Carotid artery intima-media thickness extraction based on Snake |
| Author： | WU Huihui, WU Qiong, XU Hongwei, LIU Yang, WANG Xidian, YANG Xin, DING Mingyue |
| Organization： | Key Laboratory of Image Processing and Intelligence Control, Education Ministry of China, College of Life Science and Technology, Huazhong University of Science and Technology; CT-room of the No.5 Affiliated Hospital, Zhengzhou University; State Key Laboratory for Multi-spectral Information Processing Technologies, Institute for Pattern Recognition and Artificial Intelligence (IPRAI), Huazhong University of Science and Technology |
| Abstract： | This paper proposed an automated layer extraction based on Snake (ALES) to extract ultrasound carotid intima-media thickness. After selecting the region of interest (ROI) and initializing the contour, ultrasound carotid intima was extracted by Snake while adventitia by gradient vector flow (GVF)-Snake which used gradient vector flow as its external force. Thus, the final intima and adventitia were obtained. Then, it was validated by comparing it with an improved active model approach which used a new external energy function. The validation experiment was based on ultrasound carotid images with known ground truth. The segmentation results of the two approaches were compared by calculating relative errors between ground truth and intima-media thickness (IMT) extracted by improved active model approach and ALES. It was demonstrated that the ALES approach could obtain better results to extract ultrasound carotid IMT due to its smaller relative errors. |
| Key words： | computer image processing; image segmentation; automated layer extraction based on Snake; carotid intima-media thickness |
| 发表期数： | 2014年2月第4期 |
| 引用格式： | 吴慧慧，吴琼，徐红卫，等. 基于Snake的颈动脉内中膜厚度提取[J]. 中国科技论文在线精品论文，2014，7（4）：314-321. |
0
评论数 0请您登录
暂无评论
下载全文
多维论文
