Multi-parameter acoustic imaging of uniform objects in inhomogeneous soft tissue

Area of Science:

• Medical Imaging
• Acoustics
• Optimization

Background:

• Ultrasound imaging reconstructs images from scattered sound fields.
• Accurate reconstruction is challenged by unknown object properties and background clutter.
• Frequency-domain measurements offer rich data for advanced reconstruction.

Purpose of the Study:

• To develop a robust ultrasound image reconstruction method for objects with uniform but unknown contrast in attenuation and sound speed.
• To incorporate realistic background clutter into the imaging solution.
• To improve classification performance in challenging acoustic environments.

Main Methods:

• Utilized frequency-domain scattered field measurements.
• Employed an exact scattering model for randomly located small scatterers to simulate background clutter.
• Applied a total-variation minimization-based approach with sound speed and attenuation perturbations as parameters.
• Used convex optimization for the reconstruction algorithm.
• Generated numerical data via the discretized Lippman-Schwinger equation.
• Applied a Born approximation-based statistical model for object profile reconstruction.

Main Results:

• Presented a two-dimensional ultrasound image reconstruction method.
• Evaluated performance using classification metrics.
• Compared the proposed method against minimum-l2-norm reconstruction.
• Demonstrated robust classification performance down to a signal-to-clutter ratio of less than 1 dB.

Conclusions:

• The proposed total-variation minimization-based ultrasound imaging method provides robust object classification.
• The inclusion of realistic scattering models and optimization techniques enhances performance in the presence of background clutter.
• This approach offers a significant improvement for ultrasound image analysis in complex scenarios.