BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Physics-enhanced velocimetry (PEV) for joint reconstruction and se
 gmentation of noisy Flow-MRI images - Matthew Juniper\, Professor of Depar
 tment of Engineering in University of Cambridge
DTSTART:20230123T110000Z
DTEND:20230123T120000Z
UID:TALK194608@talks.cam.ac.uk
CONTACT:Dr Song
DESCRIPTION:We formulate and solve a generalized inverse Navier–Stokes b
 oundary value problem for velocity field reconstruction and simultaneous b
 oundary segmentation of noisy flow velocity images. We use a Bayesian fram
 ework that combines CFD\, Gaussian processes\, adjoint methods\, and shape
  optimization in a unified and rigorous manner. With this framework\, we f
 ind the velocity field and flow boundaries (i.e. the digital twin) that ar
 e most likely to have produced a given noisy image. We also calculate the 
 posterior covariances of the unknown parameters and thereby deduce the unc
 ertainty in the reconstructed flow. First\, we verify this method on synth
 etic noisy images of 2-D flows. Then we apply it to experimental phase con
 trast magnetic resonance (PC-MRI) images of an axisymmetric flow at low (
 ≃6) and high (>30) SNRs. We show that this method successfully reconstru
 cts and segments the low SNR images\, producing noiseless velocity fields 
 that match the high SNR images\, despite using 27 times less data. This fr
 amework also provides additional flow information\, such as the pressure f
 ield and wall shear stress\, accurately and with known error bounds. We de
 monstrate this on a synthetic 2-D representation of the flow through an ao
 rtic aneurysm to show its relevance to medical imaging.
LOCATION:JDB - Seminar Room in the Dyson building\, Engineering department
 \, Trumpington Street\, Cambridge\, CB2 1PZ
END:VEVENT
END:VCALENDAR