 |
|
|
|
 |
 |
 |
|
 |
|
 |
|
|
段侪杰,马竟锋,张艺宝,侯凯,包尚联.能量传导模型及在医学图像分割中的应用.软件学报,2009,20(5):1106-1115 |
能量传导模型及在医学图像分割中的应用 |
Energy Conduction Model and Its Application in Medical Image Segmentation |
投稿时间:2008-08-27 修订日期:2008-12-15 |
DOI: |
中文关键词: 能量传导模型 水平集 C-V 模型 多目标分割 医学图像分割 |
英文关键词:energy conduction model level set framework C-V model multi-targets segmentation medical image segmentation |
基金项目:Supported by the National Natural Science Foundation of China under Grant Nos.10527003, 60672104 (国家自然科学基金); the National Basic Research Program of China under Grant No.2006CB705705 (国家重点基础研究发展计划(973)); the Joint Research Subject of Beijing Education Committee of China under Grant No.JD100010607 (北京市共建项目); the Beijing Municipal Natural Science Foundation of China under Grant No.3073019 (北京市自然科学基金); the Upgrading Subject of Instrument in Science and Technological Ministry of China under Grant No.2006JG1000 (科技部仪器升级改造项目) |
|
摘要点击次数: 7487 |
全文下载次数: 4593 |
中文摘要: |
提出了一种基于水平集框架的能量传导模型ECM(energy conduction model)用于对医学图像进行分割.该模型通过对图像中的灰度分布和空间中的温度场分布进行对比,有效定义了图像能量和图像能量的传导方程,并通过模拟热量传递的过程对方程进行求解.ECM模型的优点在于,它在描述图像灰度分布的全局特征的同时,有效地捕捉到图像局部区域的灰度对比度变化,因此它能够对灰度分布不均匀和含有噪声的图像进行精确分割.基于水平集函数本身的拓扑可变性,该方法还能够实现同一图像中的多目标分割.使用该方法对模拟和真实的医学图像进行了分割实验,实验结果表明了该方法的有效性和可靠性. |
英文摘要: |
This paper proposes an energy conduction model (ECM) based on the level set framework, which takes advantage of the heat conduction equation to construct the image energy. After comparing the image intensity distribution with the spatial distribution of the temperature field, an energy conduction function is defined, which well simulates the process of heat conducting. The advantage of the ECM is that it captures the global feature of an image and takes the local intensity information into account. Thus, ECM is able to accurately segment medical images with inhomogeneity and noise, as well as for the medical images with multi-targets. Synthetic and real medical images are tested with ECM, which shows its robustness and efficiency. |
HTML 下载PDF全文 查看/发表评论 下载PDF阅读器 |
|
|
|
|
|
|
 |
|
|
|
|
 |
|
 |
|
 |
|