Super-Resolution Ultrasound Microbubble Tracking via Joint Probabilistic Data Association Filter

Super-resolution (SR) ultrasound imaging can break the diffraction limit by imaging the microbubble (MB) contrast agents, which shows potential in clinical diagnosis and therapy. The microvascular structures and the corresponding hemody-namics can be established by depicting the trajectories of the MBs. However, the accurate localization and tracking of the MBs remain challenging. In the ULTRA-SR challenge, we use Otsu's method combined with a sequence of image processing steps to detect and localize MBs. Then we propose to use an adapted joint probabilistic data association filter (JPDAF) to track the localized MBs. We validate our methods on the synthetic and in-vivo datasets provided by the challenge. The qualitative experimental results demonstrate that our method can successfully detect and track MBs, depicting trajectories which highly correlate to the ground truths. Further improvements on the localization and tracking of the complex in-vivo data sets are our future work.