Phase-based block matching applied to motion estimation with unconventional beamforming strategies

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Signal Processing

Background:

• Traditional motion estimation in ultrasound elastography relies on amplitude images and complex cross-correlation functions.
• Unconventional beamforming strategies generate 2-D Radio Frequency (RF) images with axial and lateral modulations, enabling new approaches to motion analysis.

Purpose of the Study:

• To introduce a phase-based block matching method for local displacement estimation in ultrasound elastography.
• To leverage phase information from modulated RF images for analytical displacement solutions, avoiding classical cost function minimization.

Main Methods:

• Developed a phase-based block matching algorithm utilizing 2-D RF images obtained through unconventional beamforming.
• Applied an analytical solution for local displacement estimation directly to phase signals, bypassing complex cross-correlation computations.
• Evaluated the method's performance using simulated and experimental data for elastography applications.

Main Results:

• The phase-based method achieves high accuracy in axial displacement estimation, with minimal degradation even with significant signal downsampling (e.g., 3 µm mean axial error with decimation factors of 2 and 4).
• Compared to classical block matching, the proposed method requires significantly smaller image sizes (36x fewer pixels) and shorter computational times (10x faster) for comparable accuracy.
• Demonstrated a 10% reduction in error compared to methods using complex correlation function phase, and confirmed applicability to n-D signals and classical RF ultrasound images.

Conclusions:

• The phase-based block matching method provides an accurate and computationally efficient approach for motion estimation in ultrasound elastography.
• The method's robustness to signal downsampling and reduced computational demands facilitate real-time implementation.
• The technique's adaptability to various signal types and its superior performance highlight its potential for advancing ultrasound-based tissue characterization.