Non-iterative model-based inversion for low channel-count optical ultrasound imaginga).

Area of Science:

• Ultrasound imaging
• Optical ultrasound (OpUS)
• Image reconstruction algorithms

Background:

• Traditional delay-and-sum ultrasound reconstruction is suboptimal for low channel-count systems, leading to significant image artefacts.
• Model-based inversion methods can suppress artefacts but are often computationally intensive, hindering real-time application.
• The optical ultrasound (OpUS) modality utilizes very low channel counts, exacerbating artefact issues and limiting dynamic range.

Purpose of the Study:

• To develop and evaluate a non-iterative, direct model-based inversion (DMI) method for ultrasound image reconstruction.
• To address image artefacts and improve image quality in low channel-count systems, specifically optical ultrasound (OpUS).
• To enable real-time or near-real-time image reconstruction on modest computational resources.

Main Methods:

• Implementation of a direct model-based inversion (DMI) approach.
• Application of DMI to both synthetic and experimental optical ultrasound (OpUS) data.
• One-off pre-computation of system matrices to optimize reconstruction time.

Main Results:

• Substantial reduction in image artefacts and noise was achieved using the DMI method.
• Improved recovery of image amplitudes was observed in reconstructed images.
• Significantly reduced reconstruction times were realized after system matrix pre-computation, even in mildly inhomogeneous scenarios.

Conclusions:

• Direct model-based inversion (DMI) is an effective method for improving image quality in low channel-count ultrasound systems like OpUS.
• DMI offers a computationally feasible solution for reducing artefacts and noise, enhancing amplitude recovery, and speeding up reconstruction.
• The DMI approach is broadly applicable to other low channel-count imaging systems, potentially improving image quality and reducing system complexity.