Ultrafast Ultrasound Imaging as an Inverse Problem: Matrix-Free Sparse Image Reconstruction.

Area of Science:

• Medical Imaging
• Computational Ultrasound

Background:

• Conventional ultrasound (US) image reconstruction uses delay-and-sum (DAS) beamforming, which has limitations.
• Iterative techniques offer alternatives but require accurate models and significant memory for matrix coefficients.

Purpose of the Study:

• To present novel, fast, and matrix-free formulations for US imaging measurement models and their adjoints.
• To enhance ultrasound image reconstruction using sparse regularization and compressed sensing.

Main Methods:

• Developed two techniques leveraging fast, matrix-free formulations for US measurement models.
• Employed sparse regularization and compressed beamforming for image reconstruction.
• Utilized simulated and in vivo data for validation.

Main Results:

• The proposed approach is three orders of magnitude faster than non-DAS state-of-the-art methods.
• Achieved comparable or superior image quality compared to existing techniques.
• Demonstrated high-quality image restoration from reduced raw data.

Conclusions:

• The presented compressed beamforming techniques offer significant speed improvements in ultrasound imaging.
• These methods provide a computationally efficient and effective alternative for high-quality ultrasound image reconstruction.
• The approach enhances image quality while reducing data requirements and processing time.