Estimation of frequency-dependent attenuation based on parametric spectral analysis and autocorrelation-lags of demodulated echographic signal

Abstract

Two methods for mean frequency estimation that try to improve the compromise between variance on the estimation of this parameter and the analyzed tissue volume are described. In one method the centroid of the spectrum is obtained by parametric analysis, using autoregressive modeling of the simulated signal. The other method is an implementation of a mean-frequency estimator derived from the available autocorrelation logs of the simulated signal. The methods are applied to signals downconverted to baseband. The results are compared with those obtained by the zerocrossing detector, the centroid of the FFT-based spectra of RF and demodulated signals, and the color Doppler estimator. The results show good estimation precision (1.3-8%) with the present approaches for both small and large windows. The methods with which they are compared exhibit higher dispersion (4-66.6%), especially for small windows. >