New approaches in the field of ultrasound utilize the frequency content of the backscattered signal, which contains information about the scattering tissue structure and enables quantitative ultrasound imaging. The ultrasonic attenuation is a quantitative parameter that can be used to characterize tissue. Furthermore, for any scatterer size estimations, the spectral distortions caused by attenuation need to be corrected, and therefore the attenuation has to be measured in vivo to make such techniques applicable for medical imaging. There are two main techniques, spectral difference and spectral shift, used to estimate the attenuation from backscattered ultrasound data. In this work we focus on the spectral shift method, which utilizes the downshift of the center frequency of a pulse propagating through an attenuating medium. The spectral shift method is less sensitive than the spectral difference method and there has not been much work done on estimating the accuracy and precision of the attenuation estimates. The amount of downshift of the pulse depends on the square of the absolute bandwidth of the pulse and is believed to bea crucial parameter for the accuracy of the attenuation estimate. An algorithm utilizing the spectral shift method was implemented and tested on virtual and physical phantoms with different attenuation slopes. The virtual phantoms were interrogated with pulses with a wide range of center frequencies and fractional bandwidths. We found that the algorithm used to estimate the attenuation slope is unbiased but has a high variance in the estimates for situations where the amount of downshift of the pulse is small. Experiments with two transducers differing in absolute bandwidth indicated that for a fixed analysis depth higher bandwidths perform better.
【 预 览 】
附件列表
Files
Size
Format
View
Evaluation of the accuracy and precision of ultrasound attenuation slope estimates
PDF
download
878987797
028-85220240
