High-resolution 3-D microwave breast imaging using spatial prior constraints with variable strength

We investigate an inverse scattering algorithm that exploits spatial prior information for high resolution microwave breast imaging in 3D. Spatial prior information is incorporated into the distorted Born iterative method (DBIM) in the form of variable-strength soft constraints. The algorithm uses sparse matrices, which makes it computationally efficient. We test the performance of the algorithm by imaging a realistic 3D numerical breast phantom using a simulated dipole array. We demonstrate that, in contrast to conventional DBIM, the algorithm preserves the boundaries between tissue types and accurately estimates the tissue dielectric properties in the different tissue regions. improving property estimation. We also show that different levels of dielectric properties heterogeneity in the breast tissues can be captured in the reconstructed image by varying the strength of the spatial prior constraints.