Dynamic mode decomposition as a framework for denoising ultrafast power doppler images

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Biomedical Engineering

Background:

• Microvessel imaging is crucial for diagnosing and monitoring pathologies.
• Ultrafast Power Doppler (UPD) ultrasound offers portability and high temporal resolution but suffers from high noise levels due to unfocused wave transmission.

Purpose of the Study:

• To introduce a novel denoising approach for UPD imaging using Dynamic Mode Decomposition (DMD).
• To enhance the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of UPD images.

Main Methods:

• Developed a new framework linking DMD dynamic modes to ultrasound acquisitions.
• Removed temporal signals from noisy modes before UPD image calculation.
• Implemented pixel-level, energy-based adaptation of discarded modes for local noise filtering.

Main Results:

• The DMD-based denoising method was validated via simulations, phantom studies, and in vivo experiments.
• Achieved up to 26.0 dB improvement in SNR and 15.6 dB in CNR in vivo compared to standard UPD images.

Conclusions:

• The DMD-based framework significantly improves UPD image quality for enhanced vessel visualization.
• This method provides a foundation for advanced dynamic analysis in vascular ultrasound imaging.