Optimization of speed-of-sound estimation from noisy ultrasonic signals

The effects of prefiltering and the choice of time-delay estimators and statistical data reduction techniques on the precision of speed-of-sound estimation were investigated using the beam-tracking technique. It was found that prefiltering the data with an ideal 50-kHz low-pass filter improved the precision of the estimation in all cases. Echo cross-correlation had an advantage over peak detection for low signal-to-noise ratio (SNR) levels, but its advantage diminished as the signal-to-noise level improved due to filtering. The linear regression method was superior to the paired-point analysis technique under all conditions. Using the optimal set of parameters, precision on the order of 0.1% was achieved in a tissue-mimicking phantom when one beam was tracking along 75 mm in 1-mm increments.<>