Theoretical Study on High-Quality Visualization of Breast Cancer Using Microwaves

点击翻译按钮获取中文摘要

It is the purpose of this research to provide high-quality visualization algorithms of imaging strongly scattering objects such as a breast cancer. The research results are summarized as follows.1.In the analysis of frequency-domain inverse scattering problems, we employ the genetic algorithm (GA), the conjugate gradient method, and the Levenberg-Marquardt method for the reconstruction of relative permittivities of objects. On the other hand, the GA and the conjugate gradient method are used in time-domain inverse scattering problems. Both the frequency-and time-domain inverse scattering problems are reduced to nonlinear optimization problems. Then we can derive efficient visualization algorithms effective in strongly scattering objects. Note that the GA is very useful for a one-dimensional inverse scattering problem.2.In frequency-domain inverse scattering problems, the convergence speed for the multifrequency case is much faster than that for the single-frequency case. The optimal re … More gularization parameter is determined by minimizing the absolute value of the radius of curvature of the generalized cross-validation (GCV) function. This method provides successful reconstructions with the property of much faster convergence than the conventional GCV method. To accelerate the visualization algorithm, we propose a new regularization method where the regularization term is dynamically changed with the residual error in the scattered electric field.3.The effect of an incident pulse on the accuracy of image reconstruction in time-domain inverse scattering is also studied. It is confirmed that the reconstruction accuracy can be considerably improved by selecting the wave form of an incident pulse. Especially, we can obtain good reconstructions by using the monocycle pulse.It is concluded from the research results that the visualization algorithms proposed have the properties of fast-convergence and high-resolution. Then these algorithms can be applied to the imaging of a breast cancer. Less