A Combined Algorithm for High Resolution Microwave Breast Imaging Using Eigenfunction-based Prior

摘要

A new algorithm for quantitative microwave breast imaging and cancer detection is presented that is based on creating prior information for the finite element contrast source inversion using a non-iterative eigenfunction-based technique. The prior information is introduced as a numerical inhomogeneous background that modifies the contrast and contrast sources variables being inverted by the finite element contrast source inversion. It is found that this type of specialized physics-based prior, that uses the eigenfunctions of the Helmholtz operator, ensures the stability of the subsequent contrast source inversion and produces high-accuracy and high-resolution breast images. The performance of the combined algorithm is verified by simulation studies of a 2D MRI-derived anthropomorphic breast model.