• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control >Nonlocal means filter-based speckle tracking
【24h】

Nonlocal means filter-based speckle tracking

机译：非局部意味着基于滤波器的斑点跟踪

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The objective of sensorless freehand 3-D ultrasound imaging is to eliminate the need for additional tracking hardware and reduce cost and complexity. However, the accuracy of current out-of-plane pose estimation is main obstacle for full 6-degree-of-freedom (DoF) tracking. We propose a new filter-based speckle tracking framework to increase the accuracy of out-of-plane displacement estimation. In this framework, we use the displacement estimation not only for the specific speckle pattern, but for the entire image. We develop a nonlocal means (NLM) filter based on a probabilistic normal variance mixture model of ultrasound, known as Rician-inverse Gaussian (RiIG). To aggregate the local displacement estimations, Stein???s unbiased risk estimate (SURE) is used as a quality measure of the estimations. We derive an explicit analytical form of SURE for the RiIG model and use it as a weight factor. The proposed filter-based speckle tracking framework is formulated and evaluated for three commonly used noise models, including the RiIG model. The out-of-plane estimations are compared with our previously proposed model-based algorithm in a set of ex vivo experiments for different tissue types. We show that the proposed RiIG filter-based method is more accurate and less tissue-dependent than the other methods. The proposed method is also evaluated in vivo on the spines of five different subjects to assess the feasibility of a clinical application. The 6-DoF transform parameters are estimated and compared with the electromagnetic tracker measurements. The results show higher tracking accuracy for typical small lateral displacements and tilt rotations between image pairs.
机译：无传感器徒手3D超声成像的目的是消除对附加跟踪硬件的需求，并降低成本和复杂性。但是，当前平面外姿态估计的准确性是完整6自由度（DoF）跟踪的主要障碍。我们提出了一种新的基于滤波器的斑点跟踪框架，以提高平面外位移估计的准确性。在此框架中，我们不仅将位移估计用于特定的斑点图案，而且还将整个图像用于位移估计。我们基于超声的概率正态方差混合模型（称为Rician-逆高斯（RiIG））开发了非局部均值（NLM）滤波器。为了汇总局部位移估计，Stein的无偏风险估计（SURE）被用作估计的质量度量。我们得出RiIG模型的SURE的明确解析形式，并将其用作权重因子。针对三个常用的噪声模型（包括RiIG模型）制定并评估了所提出的基于滤波器的斑点跟踪框架。在一组针对不同组织类型的体外实验中，将平面外估计值与我们先前提出的基于模型的算法进行了比较。我们表明，提出的基于RiIG过滤器的方法比其他方法更准确，组织依赖性更小。还对五个不同受试者的脊柱进行了体内评估，以评估临床应用的可行性。估计6自由度变换参数，并将其与电磁跟踪器测量结果进行比较。结果表明，对于典型的较小横向位移和图像对之间的倾斜旋转，跟踪精度更高。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号