Passive Cavitation Imaging Artifact Reduction Using Data-Adaptive Spatial Filtering.

Area of Science:

• Medical Imaging
• Acoustic Imaging
• Ultrasound Technology

Background:

• Passive cavitation imaging (PCI) using clinical arrays suffers from poor axial localization due to point spread function (PSF) limitations.
• Improving source localization and image quality in PCI is crucial for accurate diagnosis and treatment monitoring.

Purpose of the Study:

• To evaluate if data-adaptive spatial filtering enhances passive cavitation imaging (PCI) beamforming performance.
• To compare spatial filtering methods against standard delay, sum, and integrate (DSI) and robust Capon beamforming (RCB).
• To achieve improved source localization and image quality without increasing computational cost.

Main Methods:

• Spatial filtering was applied as pixel-based masks to DSI or RCB beamformed images.
• Masks were generated using DSI, RCB, or coherence factors (phase/amplitude), analyzed with ROC and PR curves.
• Performance was assessed using binary classifier metrics on simulated cavitation emissions.

Main Results:

• Spatial filtering demonstrated comparable sensitivity, specificity, and AUROC to other methods (within 11% difference).
• Spatially filtered DSI methods were two orders of magnitude faster than time-domain RCB.
• Data-adaptive spatial filtering achieved similar classification performance with significantly reduced computational time.

Conclusions:

• Data-adaptive spatial filtering is a computationally efficient strategy for improving PCI beamforming.
• This approach enhances source localization and image quality in passive cavitation imaging.
• The method presents a preferable alternative for PCI applications requiring faster processing times.