Cross-correlation adjustment full-waveform inversion with source encoding in ultrasound computed tomography.

Area of Science:

• Medical Imaging
• Computational Physics
• Acoustics

Background:

• Full-waveform inversion (FWI) is a key technique in ultrasound computed tomography (USCT) for sound speed imaging.
• Conventional FWI faces challenges in stability and computational expense.
• Improving FWI robustness is crucial for broader USCT applications.

Purpose of the Study:

• To propose a novel USCT algorithm, cross-correlation adjustment FWI with source encoding (CCAFWI-SE).
• To enhance the stability and reduce the computational cost of FWI in USCT.
• To demonstrate the effectiveness of the proposed algorithm in reconstructing sound speed models.

Main Methods:

• Developed a CCAFWI-SE algorithm adjusting the gradient using cross-correlation-derived intermediate signals.
• Implemented an intermittent update strategy to decrease computational load.
• Validated the approach through numerical simulations and laboratory experiments.

Main Results:

• The CCAFWI-SE algorithm successfully converged to the ground truth sound speed model where conventional FWI failed.
• Experimental results confirmed improved stability and accurate sound speed model recovery.
• The intermittent update strategy reduced computational costs, enabling source encoding.

Conclusions:

• The proposed CCAFWI-SE algorithm significantly improves the stability and robustness of USCT reconstruction.
• This method expands the applicability of FWI in scenarios with poor initial models.
• The approach offers practical benefits by reducing computational demands and can be combined with existing techniques.