A Frequency-Hopping Technique for Solving the Cycle-Skipping Problem Encountered with Acoustic Full-Waveform Inversion

Abstract

Full-waveform inversion methods are widely used for breast imaging to provide quantitative information about the tissues such as the speed of sound. Full-waveform inversion faces a cycle-skipping problem when applied to large contrasts or large spatial domains. This problem leads to erroneous image reconstructions. To overcome this problem, low frequency measurements or good starting models are needed. In this work, we propose and test a frequency-hopping technique for our full-waveform inversion method (Contrast Source Inversion) to avoid the cycle-skipping problem. For a synthetic example, we first show the effect of cycle skipping in the reconstructed image. Next, we show for the same example that the implementation of the proposed frequency-hopping method solves this problem.