Forschungscampus
STIMULATE - Solution Centre for Image Guided Local Therapies Logo
BMBF Logo
EFRE Logo

11.01.2019 - STIMULATE-Kolloquium

Am Freitag, dem 11.01.2019 findet im Gebäude 9, Raum 2.11 ab 14.30 Uhr das STIMULATE-Kolloquium statt.

 

Vortragender:

Prof. Dr.-Ing. habil. Andreas Maier, Universität Erlangen-Nürnberg, Lehrstuhl für Mustererkennung (Informatik 5)

 

Vortragstitel: 

"Towards Learning of Efficient Solutions for Image Reconstruction"

 

Abstract:

The presentation will start with with a short overview on the activities in pattern recognition and image reconstruction in Erlangen. This is followed by the introduction of a mathematical framework for learning an efficient solution for the limited angle reconstruction problem, that can typically not be solved using the well known Radon Inversion. Being discrete from end to end, the approach is also able to learn compensation strategies such as discretisation intrinsically given the continuous solution for the problem. Furthermore, we show relations to other imaging geometries such as cone-beam geometries and that we are technically able to optimally reconstruct arbitrary scan trajectories.

Furthermore, we demonstrate that our approach which elegantly combines traditional reconstruction theory with deep learning, is also optimal in the sense of maximal learning error and provide theoretic and practical arguments why this approach is superior to image to image artefact reduction using, e.g. the well-known U-net. 

The presentation ends with a perspective on efficient inverse problem schemes for applications in other imaging modalities such as cone-beam to parallel-beam conversion or other imaging problems.

CV:

Prof. Dr. Andreas Maier was born on 26th of November 1980 in Erlangen. He studied Computer Science, graduated in 2005, and received his PhD in 2009. From 2005 to 2009 he was working at the Pattern Recognition Lab at the Computer Science Department of the University of Erlangen-Nuremberg. His major research subject was medical signal processing in speech data. In this period, he developed the first online speech intelligibility assessment tool - PEAKS - that has been used to analyze over 4.000 patient and control subjects so far.
From 2009 to 2010, he started working on flat-panel C-arm CT as post-doctoral fellow at the Radiological Sciences Laboratory in the Department of Radiology at the Stanford University. From 2011 to 2012 he joined Siemens Healthcare as innovation project manager and was responsible for reconstruction topics in the Angiography and X-ray business unit. 
In 2012, he returned the University of Erlangen-Nuremberg as head of the Medical Reconstruction Group at the Pattern Recognition lab. In 2015 he became professor and head of the Pattern Recognition Lab.

 

Interessierte sind herzlich eingeladen.

 

zur Kolloquiumsübersicht