BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Learning iterative reconstruction for high resolution photoacousti
 c tomography - Andreas Hauptmann (University College London)
DTSTART:20171031T172000Z
DTEND:20171031T181000Z
UID:TALK94129@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Recent advances in deep learning for tomographic reconstructio
 ns have shown great potential to create accurate and high quality images w
 ith a considerable speed-up. In this work we present a deep neural network
  that is specifically designed to provide high resolution 3D images from r
 estricted photoacoustic measurements. The network is designed to represent
  an iterative scheme and incorporates gradient information of the data fit
  to compensate for limited view artefacts. Due to the high complexity of t
 he photoacoustic forward operator\, we separate training and computation o
 f the gradient information. A suitable prior for the desired image structu
 res is learned as part of the training. The resulting network is trained a
 nd tested on a set of segmented vessels from lung CT scans and then applie
 d to  in-vivo photoacoustic measurement data.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR