Two-Dimensional Diffraction Tomography with Noise Suppression Using Principal Component Analysis

Chunxia Yang,Jian Zhang,Meisong Tong
DOI: https://doi.org/10.1109/COMPEM.2019.8778870
2019-01-01
Abstract:In inverse scattering problems, the field inside the imaging domain is unattainable. Moreover, the involved governing equations are nonlinear and the solutions are inherently non-unique. Particularly, when the measurement data are contaminated by noise, solving inverse scattering problems is even more challenging. Therefore, in this paper, we take advantage of Principal Component Analysis (PCA) method to filter the noise hidden in the scattered field data. In PCA algorithm, only the first few principal components are kept, while the other components, which may contain noise information, are ignored. Numerical example is included to illustrate the improvement of imaging quality benefitting from PCA.
What problem does this paper attempt to address?