Super-Resolved Micro-Vascular Imaging Using Microbubbles and Machine Learning
Christensen-Jeffries, Kirsten ; Schirmer, Markus D ; Browning, Richard ; Tang, Meng-Xing ; Dunsby, Christopher ; Aljabar, Paul ; Eckersley, Robert John
Recent developments in sub-diffraction ultrasound (US) imaging using clinical US systems has shown the potential to resolve structures on the micrometer scale using microbubbles (MBs). These rely on user-defined thresholds for MB identification making their clinical application challenging. Here, an automated post-processing algorithm based on k-means clustering has been developed to identify noise, individual and multiple MB in vivo without user interaction. This method has the potential to non-invasively image in real-time pathological or therapeutic changes in the micro-vasculature at centimeter depths in a clinical setting.