Spectral Analysis Methods with Continuous Wavelet Transform to Characterize Media Periodicity Using Backscattered and Simulated Ultrasound Signals

Abstract

This study aimed at applying spectral methods and CWT in the periodicity characterization in simulated/backscattered A-mode ultrasound signals. The experimental setup used a transducer (7.5 MHz) on a pulse-echo configuration. A copper-rods phantom modeled a periodic scatterers pattern (diameter = 0.5 mm, spacing = 2.5 mm). The transducer was translated parallel to the phantom to collect signals. A-mode 7.5-MHz signals were also simulated using a numerical model with contributions from regular (MSS = 1 mm) and diffuse scatterers, with several levels of noises and jitters (variability on the expected scatterer positions). Experimental and simulated signals were processed with the spectral autocorrelation (SAC), singular spectrum analysis (SSA) and the quadratic transformation (Simon’s method) for MSS estimation, as well as the continuous wavelet transform with two mother-wavelets. MSS distributions are not similar in many noise-jitter situations, or when comparing the performance between two methods (p < 0.05). SAC was able to better characterize scatterer spacing. SIMON and CWT presented modes near 2.5 mm and 3.5 mm (the MSS and the diagonal distance of two rods). SSA may detect rods thicknesses (0.5 mm). CWT would fit for locating scatterer position.