Spatially segmented SVD clutter filtering in cardiac blood flow imaging with diverging waves

Area of Science:

• Medical Imaging
• Ultrasound Technology
• Cardiovascular Imaging

Background:

• Ultrafast ultrasound imaging visualizes rapid cardiac blood flow dynamics.
• Singular value decomposition (SVD) filters are effective for ultrafast blood flow imaging, outperforming conventional filters in specific applications.
• Implementing SVD filters in cardiac imaging is challenging due to complex, variable tissue characteristics.

Purpose of the Study:

• To develop and evaluate a novel method for improving SVD filter implementation in ultrafast cardiac ultrasound imaging.
• To address the challenges of complex spatial and temporal tissue variations in cardiac imaging.
• To enhance the accuracy of tissue clutter rejection and blood signal visualization.

Main Methods:

• A spatially segmented approach was developed, excluding the proximal image portion and dividing the field of view into overlapped segments.
• Singular value decomposition (SVD) filtering with automatic cut-off singular vector order selection was applied to each segment.
• Auto-thresholding was based on spatial singular vector coherence to delineate tissue, blood, and noise subspaces.
• Filtered blood flow signals were reconstructed, combined, and processed using Doppler to create blood flow images.

Main Results:

• The spatially segmented approach improved the separation of tissue and blood subsets in the spatial similarity matrix.
• Automatic thresholding accuracy increased from 78% to 98.7% in investigated cases.
• Post-filter power of blood signals improved by an average of over 10 dB during a cardiac cycle.
• Tissue clutter rejection was more effective in cardiac ultrafast imaging compared to using the full field of view.

Conclusions:

• Spatially segmented SVD filtering is an effective method for improving ultrafast cardiac ultrasound imaging.
• The proposed method enhances automatic thresholding and tissue clutter rejection, leading to better blood flow visualization.
• This technique offers a significant advancement for cardiac ultrafast imaging applications.