by Deep Learning
Milecki L; Poree J; Belgharbi H; Bourquin C; Damseh R; Delafontaine-Martel P; Lesage F; Gasse M; Provost J, 2021. A Deep Learning Framework for Spatiotemporal Ultrasound Localization Microscopy. IEEE Trans Med Imaging 40(5):1428-1437
Ultrasound localization microscopy allows imaging the blood flow in organs in a fast, non-invasive and non-ionizing way. Based on the detection of microbubbles injected into the bloodstream, it is possible to obtain ultra-resolved angiograms of organs. The classical method seeks to detect isolated microbubbles and therefore requires that they be injected in small quantities. In order to image the entire blood network, it is then necessary to wait until even the smallest vessels have been covered by microbubbles. The use of neural networks could allow to detect the trajectories of microbubbles in higher concentration and to reduce the acquisition time.