A Regularization Approach for Frequency-Dependent Attenuation Measurement

Ultrasound attenuation is frequency dependent. The non-linear frequency dependence can be characterized by using a power-law model, <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\alpha(f)=af^{b}$</tex>. The frequency dependence of attenuation is correlated with underlying tissue structure, and can be used as clinically useful biomarker of different conditions. We previously proposed a total variation regularization method to reconstruct ultrasound attenuation, with a simplifying assumption of linear frequency dependence. In this paper, we extend the method to estimate the attenuation as a function of frequency. We fit a power law model to characterize the relationship between attenuation and frequency. We apply the proposed method for a broadband attenuation characterization (2–20 MHz) of a Zerdine™-based tissue-mimicking phantom and compare the performance of the proposed method with a baseline reference phantom method. We demonstrated that the proposed method can reliably characterize the frequency dependence of attenuation while reducing the length of the time-gated window to half compared to the reference phantom method.