Adaptive blood velocity estimation in medical ultrasound

This paper investigates the use of data-adaptive spectral estimation techniques for blood velocity estimation in medical ultrasound. Current commercial systems are based on the averaged periodogram, which requires a large observation window to give sufficient spectral resolution. Herein, we propose a novel data-adaptive method to form the blood velocity spectral estimate.

The method is evaluated using realistic field II simulations for both steady and unsteady flow. The latter representing the femoral artery with strong tissue interference. The method is compared to the averaged periodogram and a Capon-based estimator. The simulations indicate that the proposed method offer a significant performance gain, suggesting that the frame-rate may be increased dramatically by using adaptive spectral estimation techniques.