Abstract

Ultrasound tomography is an emerging clinical imaging modality that recently gained US FDA approval for adjunctive breast cancer screening. Several technical advances over the past two decades have enabled waveform inversion-based image reconstruction on clinical systems. In a previous talk, Dr. Nebojsa Duric described the current state-of-the-art in ultrasound tomography as well as future directions for this technology. In this talk, Dr. Rehman Ali will continue from the previous talk to describe in greater detail the various mathematical and technical aspects of waveform inversion (e.g., backprojection, source estimation, cycle skipping) as well as new modeling strategies towards 3D reconstructions (e.g., iterative solvers and approximations to the 3D Helmholtz equation). The intention is to demonstrate the actual algorithms used on real in-vivo breast ultrasound tomography data. As a result, this presentation will also involve many practical considerations (e.g., data preprocessing) when translating these mathematical concepts to a real hardware-software implementation of ultrasound tomography. If time permits, there may be a live MATLAB demo of the reconstruction process.   Authors: Rehman Ali (University of Rochester Medical Center), Trevor Mitcham (University of Rochester Medical Center), Nebojsa Duric (University of Rochester Medical Center)