Separation of multipath transmission signals for quantitative ultrasound measurement at bone - A comparison study

For separating multipath transmission signals from quantitative ultrasound measurements at bone several approaches have been proposed in the last years. Here, the performances of the two fastest algorithms are compared, which both estimate the parameters of the transfer functions of the sound paths: the modified least squares Prony method (MLSP) and the space alternating generalized expectation maximization algorithm (SAGE). It is shown, that for good signal-to-noise ratios (SNR) of 100 dB and broadly overlapping waves the MLSP algorithm yields better results, but with decreasing SNR the robustness to noise of the SAGE algorithm prevails.