P2F-8 Performance Comparison of Attenuation Estimation Techniques

Ultrasound attenuation measurements in soft tissue is important both for the clinical diagnosis of various pathological conditions as well as for further analysis of ultrasound B-mode images artifacts such as shadowing and enhancement. In this paper, we present a new method for estimating local attenuation coefficients using spectral cross-correlation between consecutive power spectra obtained from the backscattered radiofrequency echo signals at different depths. Since the spectral cross-correlation method estimates the spectral shift by comparing the entire power spectra, it is more robust and stable to the noise artifacts in the backscattered radiofrequency signals. We compare the attenuation estimation performance of spectral cross-correlation to the centroid downshift, and reference phantom methods reported in the literature. Ultrasound simulations that incorporate variations in the transducer dependent parameters and backscatter properties of tissue are utilized to compare the accuracy and precision of each method. Simulation results demonstrate that the estimation accuracy of the reference phantom and spectral cross-correlation methods are better than the centroid downshift method in uniform backscatter regions. However, as backscatter properties change along the beam propagation path, the reference phantom method fails to estimate attenuation coefficients at the boundaries where the backscatter changes occur while the spectral cross-correlation and centroid downshift are fairly independent of backscatter variations. Experimental results using tissue mimicking phantoms also illustrate that the spectral cross-correlation method is more robust towards changes in tissue backscatter properties