Impact of correlation level of eigenvectors on Mean Scatterer Spacing estimation Guillermo Cortela

Abstract

Tissues exhibiting quasi-periodic structures are modeled as a collection of diffuse and coherent scatterers. Coherent components are quantified by mean scatterer spacing (MSS) and are closely related to tissue structure. MSS is a quantitative ultrasound parameter useful for characterizing quasi-periodic tissues and must be determined accurately. The study aims at determining the impact of correlation level (CL) of eigenvectors on MSS estimated by Singular Spectrum Analysis (SSA), compared to other spectral methods, of a periodic distribution of scatterers using backscattered ultrasound signals. Five transducers (1, 2.25, 3.5, 5 and 7.5 MHz) were used on a pulse-echo configuration. Transducers were moved parallel to the phantoms (matrix of cylindrical copper bars) to collect signals for processing with the spectral autocorrelation, SSA and quadratic transformation for MSS estimation. The sensitivity of the SSA was analyzed for CL values of 0.5, 0.6, 0.7 and 0.8. It was evidenced that as the level of correlation decreases, the sensitivity of the methods decreases. The extreme sensitivity values of the SSA technique are 0.05 (CL =0.8,7.5 MHz) and 5.5 (CL =0.5,1 MHz). For these extreme values, the MSS estimation error was 0.8% and 18% respectively, for the SSA technique.