Interpolation methods for time-delay estimation using cross-correlation method for blood velocity measurement.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Doppler ultrasound's cross-correlation method (CCM) estimates blood flow velocity via pulse-to-pulse echo time delay.
• Low sampling frequencies reduce computational load but can introduce bias in peak correlation estimation.
• Existing parabolic-fit interpolation methods show limitations at low sampling rates relative to ultrasound center frequency.

Purpose of the Study:

• To enhance the accuracy of blood flow velocity estimation in Doppler ultrasound.
• To address the bias introduced by traditional interpolation methods at low sampling rates.
• To evaluate novel interpolation techniques for improved time delay estimation in CCM.

Main Methods:

• Developed four new interpolation methods: bias-compensated parabolic, combined parabolic and linear filter, complex envelope parabolic, and matched filter interpolation.
• Analyzed methods using computer-simulated Doppler signals and patient-recorded RF-signals.
• Investigated performance under conditions where time delay exceeds the inverse of the center frequency (1/f(0)).

Main Results:

• All four proposed methods demonstrated higher accuracy than the standard parabolic-fit method in simulations.
• Estimation accuracy improved from a worst-case 1.25% of 1/f(0) to under 0.5% of 1/f(0) with the new methods.
• Matched filter interpolation exhibited superior performance, especially at low signal-to-noise ratios (SNR), confirmed by both simulation and experimentation.

Conclusions:

• The proposed interpolation techniques significantly improve blood flow velocity estimation accuracy in Doppler ultrasound.
• Bias-compensated parabolic and matched filter interpolation offer robust alternatives to traditional methods, particularly in low SNR scenarios.
• These advancements hold potential for more precise non-invasive blood flow assessment.