A comprehensive theoretical performance comparison of the wideband maximum-likelihood (WMLE) and cross-correlation strategies, previously proposed and evaluated for the estimation of blood velocity using ultrasound is presented. It is based on evaluation of the bias, local and global accuracy, and signal-to-noise ratio (SNR) performance. The results show that the intervening medium does not bias either wideband estimator, due to the effect of tracking the scattering target. The presence of intervening tissue actually improves the global accuracy of both wideband estimators, without a significant change in the local accuracy of either wideband estimator. After the transmission of P pulses, a comparison of the performance of the two strategies shows that the cross-correlation estimator requires P2 correlations to achieve performance similar to that the WMLE with P operations. In addition, the WMLE can increase the effective SNR in comparison with cross correlation.
|Original language||English (US)|
|Number of pages||14|
|Journal||IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control|
|State||Published - Nov 1992|
ASJC Scopus subject areas
- Electrical and Electronic Engineering
- Acoustics and Ultrasonics