Statistics of scatterer property estimates

Quantitative ultrasound (QUS) techniques are based on providing parameter estimates from ultrasound backscattered signals that can be related to different properties of the tissue. Parameter estimates based on analyzing the spectrum of the ultrasound backscattered signal or the amplitude distribution of the envelope require a certain number of samples to produce meaningful estimates in terms of bias and variance of estimates. For example, calculation of the periodogram is used to approximate the true backscattered power spectrum of the ultrasound signal. Typically, the larger the samples size the better the periodogram represents the backscattered power spectrum and the better the bias and variance of QUS estimates. Analysis of the statistics of parameter estimation for spectral-based parameters and envelope statistics will allow the tradeoff between sample size and estimate bias and variance to be quantified. This chapter discusses the statistics of QUS property estimation, the effects of estimate bias and variance on the resolution of QUS parameter imaging, and techniques to reduce the variance of different QUS property estimates.