Filter-based compounded delay estimation with application to strain imaging.

Area of Science:

• Medical imaging
• Ultrasound technology
• Signal processing

Background:

• Speckle and phase fluctuations in ultrasound echoes create correlated errors in motion estimation.
• Compounding multiple decorrelated views (looks) is a known method to mitigate these estimation errors.

Purpose of the Study:

• To introduce a novel filter bank approach for creating decorrelated ultrasound looks.
• To evaluate the effectiveness of this filter bank compounding method for motion estimation accuracy.

Main Methods:

• Utilizing a filter bank to generate multiple decorrelated ultrasound "looks" from lateral filtering.
• Employing Gabor filters to analyze signal decorrelation with Gaussian apodization.
• Implementing phase-based estimators for displacement and strain calculations.
• Simulating the method on a deforming elastic phantom model.

Main Results:

• Lateral filtering preserves delay estimation accuracy within sub-bands while inducing decorrelation between them.
• Simulations show reduced root mean square error (RMSE) for displacement and strain using the compounded method compared to conventional approaches.
• Experimental data on an elastic phantom demonstrated improved strain signal-to-noise ratio (SNR).

Conclusions:

• The proposed filter bank compounding method effectively reduces motion estimation errors in ultrasound imaging.
• This technique offers improved accuracy for displacement and strain estimation, particularly in scenarios with tissue deformation.
• The findings suggest a promising approach for enhancing quantitative ultrasound elastography.