RECONSTRUCTION OF MICROWAVE ABSORPTION OF MULTIPLE TUMORS IN HETEROGENEOUS TISSUE FOR MICROWAVE-INDUCED THERMO-ACOUSTIC TOMOGRAPHY

Abstract

Time-of-∞ight (TOF) has been used to estimate sound velocity (SV) distribution of heterogeneous tissue to relieve the efiect of acoustic heterogeneity in microwave-induced thermo-acoustic tomography (MITAT). Accurately picking the TOFs is signiflcantly important to ensure high accuracy SV images, which greatly help to reconstruct the microwave absorption distribution accurately. However, current methods for picking the TOFs are designed for single source case. For breast tumor detection in MITAT, these methods become inefiective or even fail at the situation where multiple tumors are embedded in a normal breast tissue. In order to accurately reconstruct the microwave absorption properties of tumors in heterogeneous tissue in MITAT, an e-cient method for picking tumors' TOFs is proposed. Combining the advantages of the wavelet transform and Akaike information criterion (AIC), the proposed method introduces a concept of separate extraction of TOFs. It can e-ciently and accurately pick the TOFs of difierent tumors from the measured data in MITAT. Using the TOFs picked by the proposed method can e-ciently help to reduce the efiect of acoustic heterogeneity and greatly improve the accuracy and the image contrast of reconstructed microwave absorption properties. Some numerical simulations are given to demonstrate the efiectiveness and feasibility of the proposed method in this paper.