Golay-Net: Deep learning-based Golay coded excitation for ultrasound imaging.

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Artificial Intelligence

Background:

• Conventional ultrasound imaging uses short pulses, limited by signal attenuation and reduced penetration.
• Golay-coded excitation improves SNR and depth but halves the frame rate due to dual transmissions.

Purpose of the Study:

• To develop a deep learning framework to overcome the frame rate limitation of Golay-coded ultrasound excitation.
• To enhance ultrasound imaging depth and SNR without compromising real-time performance.

Main Methods:

• A novel deep learning framework, Golay-Net, based on a 1-D U-Net architecture was developed.
• Golay-Net synthesizes the echo signal for Code B from the echo signal of Code A by modifying range sidelobe phases.

Main Results:

• Golay-Net successfully synthesized code B-related echo signals with high fidelity.
• Reconstructed ultrasound images showed enhanced SNR and imaging depth.
• The proposed method maintained the original frame rate.

Conclusions:

• Golay-Net offers a solution to the trade-off between imaging performance and frame rate in Golay-coded ultrasound.
• This deep learning approach significantly improves ultrasound diagnostic capabilities.