Time-varying autoregressive spectral estimation for ultrasoundattenuation in tissue characterization

摘要

In the field of biological tissue characterization, fundamentalnacoustic attenuation properties have been demonstrated to havendiagnostic importance. Attenuation caused by scattering and absorptionnshifts the instantaneous spectrum to the lower frequencies. Due to thentime-dependence of the spectrum, the attenuation phenomenon is antime-variant process. This downward shift may be evaluated either by thenmaximum energy frequency of the spectrum or by the center frequency. Innorder to improve, in strongly attenuating media, the results given bynthe short-time Fourier analysis and the short-time parametric analysis,nwe propose two approaches adapted to this time-variant process: annadaptive method and a time-varying method. Signals backscattered by annhomogeneous medium of scatterers are modeled by a computer algorithmnwith attenuation values ranging from 1 to 5 dB/cm MHz and a 45 MHzntransducer center frequency. Under these conditions, the preliminarynresults obtained with the proposed time-variant methods, compared withnthe classical short-time Fourier analysis and the short-timenauto-regressive (AR) analysis, are superior in terms of standardndeviation (SD) of the attenuation coefficient estimate. This study,nbased on nonstationary AR spectral estimation, promises encouragingnperspectives for in vitro and in vivo applications both in weakly andnhighly attenuating media