Ultrasound elastography based on multiscale estimations of regularized displacement fields
Ultrasound elastography based on multiscale estimations of regularized displacement fields
复制标题
DOI：
10.1109/tmi.2003.822825
复制
发表时间：
2004-02-01
影响因子：
10.6
通讯作者：
Herment, A
中科院分区：
文献类型：
--
作者：
Pellot-Barakat, C;Frouin, F;Herment, A
弹性成像是基于局部组织变形的测量。本文提出的超声弹性成像的方法依赖于密集的位移场的估计由粗到细的能量函数的最小化，结合约束的回波幅度和位移场的连续性的守恒。多尺度优化方案提出了几个特点，旨在改善和加速收敛的最小化过程。这包括非正则化初始化的粗分辨率和自适应配置空间的使用。使用从组织样体模材料获得的数据调整能量模型和优化的参数。随后获得这些参数的弹性图像从正常的体内乳腺组织。在运动场估计中引入平滑度约束有助于解决由于不相干运动引起的模糊性，从而导致弹性图比从基于相关的技术获得的弹性图更少地被去相关噪声降级。
Elasticity imaging is based on the measurements of local tissue deformation. The approach to ultrasound elasticity imaging presented in this paper relies on the estimation of dense displacement fields by a coarse-to-fine minimization of an energy function that combines constraints of conservation of echo amplitude and displacement field continuity. The multiscale optimization scheme presents several characteristics aimed at improving and accelerating the convergence of the minimization process. This includes the nonregularized initialization at the coarsest resolution and the use of adaptive configuration spaces. Parameters of the energy model and optimization were adjusted using data obtained from a tissue-like phantom material. Elasticity images from normal in vivo breast tissue were subsequently obtained with these parameters. Introducing a smoothness constraint into motion field estimation helped solve ambiguities due to incoherent motion, leading to elastograms less degraded by decorrelation noise than the ones obtained from correlation-based techniques.