Magnetic induction tomography: evaluation of the point spread function and analysis of resolution and image distortion
Magnetic induction tomography: evaluation of the point spread function and analysis of resolution and image distortion
复制标题
磁感应断层扫描：点扩散函数的评估以及分辨率和图像失真的分析
DOI：
10.1088/0967-3334/28/7/s24
复制
发表时间：
2007
影响因子：
3.2
通讯作者：
H. Scharfetter
中科院分区：
文献类型：
--
作者：
R. Merwa;H. Scharfetter
磁感应断层成像（MIT）是一种低分辨率成像模式，用于重建目标物体的无源电特性的变化。对于成像系统来说，给出图像质量的预测是非常重要的。最大分辨率和位置的不均匀性的正确性是主要的利益。此外，对于特定噪声水平可以检测到的最小对象是图像的诊断价值的标准。MIT图像的特性取决于物体内部的位置、电导率分布，当然还取决于激励线圈和接收线圈的位置和数量。一般不能进行定量陈述，但对于选定的问题，预测图像质量是可行的。对于电阻抗断层成像（EIT），已经仔细研究了图像质量的理论极限，并且有必要对MIT进行全面的分析。因此，一个简单的分析分辨率，尺寸和位置的不均匀性进行了评估的点扩展函数（PSF）的装置。与EIT类似，PSF强烈依赖于位置，在对象的中心显示最宽的分布。根据增加的测量噪声增加正则化的量，PSF加宽并且其中心朝向对象的边界移动。分辨率与PSF的宽度成正比，并且当从对象的中心向边界移动时增加，并且随着噪声的增加而减小。
Magnetic induction tomography (MIT) is a low-resolution imaging modality used for reconstructing the changes of the passive electrical properties in a target object. For an imaging system, it is very important to give forecasts about the image quality. Both the maximum resolution and the correctness of the location of the inhomogeneities are of major interest. Furthermore, the smallest object which can be detected for a certain noise level is a criterion for the diagnostic value of an image. The properties of an MIT image are dependent on the position inside the object, the conductivity distribution and of course on the location and the number of excitation coils and receiving coils. Quantitative statements cannot be made in general but it is feasible to predict the image quality for a selected problem. For electrical impedance tomography (EIT), the theoretical limits of image quality have been studied carefully and a comprehensive analysis for MIT is necessary. Thus, a simplified analysis on resolution, dimensions and location of an inhomogeneity was carried out by means of an evaluation of the point spread function (PSF). In analogy to EIT the PSF depends strongly on the location, showing the broadest distribution in the centre of the object. Increasing the amount of regularization according to increasing measurement noise, the PSF broadens and its centre is shifted towards the borders of the object. The resolution is indirectly proportional to the width of the PSF and increases when moving from the centre towards the border of the object and decreases with increasing noise.