Contrast-enhanced microwave imaging of breast tumors using sparsity regularization

Abstract

We propose a technique for microwave imaging of breast tumors using contrast agents to influence the tumor dielectric properties. Scattered field data obtained before and after the agent alters the tumor properties are used to formulate a microwave inverse scattering problem whose solution is the change in dielectric properties introduced by contrast agent. A relatively small number of voxels are assumed to be influenced by the contrast agent, so the change in dielectric properties is obtained using a regularizer that encourages sparse solutions. A computational study using 2D numerical breast phantoms indicates tumors are accurately localized by this method and suggests this is a promising approach to tumor localization even in dense breast tissue.