After finishing my master’s degree in Mathematics in Science and Technology at the TU Vienna (Austria) I joined the Centre for Medical Image Computing in January 2015. Within the GIFT-Surg project, my research deals with the reconstruction of volumetric fetal MRI from 2D slices. These 3D models can provide valuable information on the specific anatomy and pathology of the fetus in order to reduce risks during complex fetal surgery by enabling and facilitating pre-surgical planning and disease diagnosis.
Prior to the beginning of my doctorate at UCL, I took part in several research projects where I was able to acquire knowledge in different fields of applied mathematics. My bachelor’s thesis was dedicated to the Adaptive Boundary Element Method and my master’s thesis dealt with the modelling of the circulatory system being undertaken as a joint-work with the AIT Austrian Institute of Technology in Vienna (Austria). Additionally, I got the possibility to contribute to the development of an adjustable throttle system for implantable infusion pumps by means of flow simulations of fluids in porous ceramics in the course of an internship at the University of Applied Sciences in Luebeck (Germany). Moreover, I was an exchange student at the Universidad de Alicante (Spain), and a visiting student at the Universidad Politécnica de Madrid (Spain), the Politecnico di Milano (Italy) and the Universität zu Lübeck (Germany).
Based on these experiences I highly appreciate having the opportunity to work in a diverse team structure with different backgrounds and sound expertise here at UCL in order to conduct my research in medical image computing and image reconstruction under the supervision of Prof. Sebastien Ourselin, Dr. Tom Vercauteren, Dr. David Atkinson and Prof. Anna David.
Despite recent advances in hardware and software, magnetic resonance imaging (MRI) remains a relatively slow imaging technique and its application difficult in the context of motion. In situations where motion cannot be avoided like in cardiac, abdominal or fetal imaging, typically, ultra-fast 2D MRI is applied due to its ability to largely "freeze" motion. Sequential acquisitions yield an image "stack" of single, thick 2D slices used for further clinical assessment. However, with motion commonly occurring between each slice acquisition, the resulting stack of thick slices offers an anatomical visualization with overall only limited geometric integrity where adjacent slices do not necessarily visualize adjacent anatomy. Moreover, the large slice thickness may hide small tissue structures and is, thus, limited in capturing fine anatomical details. In further consequence, a reliable and fine-grained diagnosis based on such motion-corrupted, thick-slice MRI data may not be possible.
My doctoral research is about the development of a volumetric MRI reconstruction framework to reconstruct geometrically consistent, high-resolution 3D volumes from motion-corrupted, possibly sparse, low-resolution 2D MR slices by using, so-called, "Super-Resolution" techniques. Instead of acquiring a high-resolution 3D MR volume directly (which cannot be achieved due to motion), the aim is its estimation to facilitate disease diagnosis and pre-surgical management by using the information of multiple, possibly motion-corrupted, low-resolution 2D slices. My focus relies on techniques applicable to a broad field of clinical applications where the use of thick-slice MR data is state-of-the-art in current clinical practice including fetal MRI and upper abdominal MRI.
Journal Articles and Conference Papers
Ebner, M., Chouhan, M., Patel, P. A., Atkinson, D., Amin, Z., Read, S., Punwani, S., Taylor, S., Vercauteren, T., and Ourselin, S. (2017). Point-Spread-Function-Aware Slice-to-Volume Registration: Application to Upper Abdominal MRI Super-Resolution. In Reconstruction, Segmentation, and Analysis of Medical Images. RAMBO 2016. Lecture Notes in Computer Science, vol 10129. Springer International Publishing, pp. 3–13. [pdf] [slides]
Ebner, M., Aertsen, M., Pratt, R., David, A. L., Pandya, P. P., Deprest, J., Kendall, G., Klusmann, M., Humphries, P., Hewitt, R., Atkinson, D., Vercauteren, T., and Ourselin, S. (2016). Volumetric MRI Reconstruction of Trachea in Fetuses with Complex Fetal Neck Masses. In Proceedings of the Institute of Physics and Engineering in Medicine (IPEM). Fetal, Neonatal and Paediatric MR Imaging: Techniques and Applications. [pdf]
Ebner, M., Hametner, B., Parragh, S., and Wassertheurer, S. (2015). Reservoir Wave Paradigm: An Implementation and Sensitivity Analysis. SNE Simulation Notes Europe, 25(3–4), 151–156. [pdf]
Aurada, M., Ebner, M., Feischl, M., Ferraz-Leite, S., Führer, T., Goldenits, P., Karkulik, M., Mayr, M., and Praetorius, D. (2014). HILBERT—a MATLAB implementation of adaptive 2D-BEM. Numerical Algorithms, 67(1), 1-32. [pdf]
Ebner, M., Mutlu, Y. S., Nestler, B., and Glatt, E. (2014). Flow Optimisation through Porous Ceramic Throttle. Student Conference Medical Engineering Science 2014 Proceedings, 75-78, Grin Verlag Gmbh. [pdf]
Ebner, M., and Winkler, S. N. (2014). Comparison of Finite Difference Method and Random Walk Method in ARGESIM Benchmark C19 "Pollution in Groundwater Flow". Simulation Notes Europe, 24(1), 51–54. [pdf]
Ebner, M. (2014). Reservoir Theory and its Application on Peripheral Arteries. Master's Thesis, Vienna University of Technology. [pdf]
Ebner, M. (2012). A-posteriori Fehlerschätzer für die Symm'sche Integralgleichung. Bachelor's Thesis, Vienna University of Technology. [pdf]