Reducing the Effects of Inhomogeneous Density with a Clustering Method in Microwave Induced Thermo-Acoustic Tomography

Abstract

Microwave induced thermo-acoustic tomography (MITAT) is a hybrid imaging technology which has drawn widespread attention. The inhomogeneous acoustic impedance that depends on velocity of acoustic and density will affect the imaging quality in MITAT. In this paper, we adopt a clustering method to reconstruct the distribution of inhomogeneous density instead of an assumption of a constant density. The clustering method divide the region of interest (RDI) into several parts and these parts represent different densities which mean different tissues. Experiment demonstrates that adopting a clustering method., the density distribution is more accurate which results in better imaging quality. This research indicates that reducing the effects of inhomogeneous density will improve the imaging quality.