Reconstruction of Flow Velocity inside Vessels by Tracking Single Microbubbles with an MCMC Data Association Algorithm

In this contribution, we propose a method to track single microbubbles. The aim of this method is the reconstruction of microvessels. An experiment with a vessel phantom was set up. The flow rate of a suspension (concentration 20/ml) of cyanoacrylate microbubbles was adjusted by a syringe pump. B-Mode images were acquired with a Vevo 2100 small animal imaging system (Visualsonics) at 255 fps. The positions of single microbubbles were identified from B-Mode images by a median filter based foreground/background separation algorithm. These positions were used by a novel Markov chain monte carlo data association (MCMCDA) algorithm to estimate the microbubbles' velocity. The mean values (8.7, 4.17, and 2.63 mm/s) of estimated single mircobubble speeds were in good agreement with the adjusted flow speed (8.8, 3.5 and 1.8 mm/s).