Alumni

Current research / employment

Currently working at the Universität Saarland.

Abstract of IMPRS PhD project:

Optical imaging is a powerful tool for monitoring physical properties of biological tissue. The use of fluorescent markers enables a selective contrast in optical imaging. Mainly there are three factors determining the resolution of optical imaging systems: light transport modeling (forward model), the image reconstruction modeling (inverse model), the instrumental setup. The main scope of my PhD work is to present an analytical light propagation method for a collimated broad excitation beam and a linear image reconstruction algorithm for under-determined inverse problem.

The coupled diffusion equations for excitation and emission light propagation are linear elliptic equations, which can be analytically solved as boundary value problem. Under those conditions exact solutions can be obtained for easy geometries.

A finite mesh was applied to the medium and point sources were considered to be located at the centers of mesh elements. Non-negative bounds were introduced for the Linear optimization method (L1). General least squares optimization method (L2) in combination with iterative search algorithm was implemented for over-determined inverse problems. The forward model is based on the diffusion approximation of the radiative transport equation and solves the excitation light as well as the fluorescence emission transports.

In the experimet three different continuous-wave light sources (different beam sizes and collimation qualities) were used. Various profiles and diameters of excitation beams have been modeled.

Reconstructed images of single and multiple fluorescent point sources are shown for over- and under-determined inverse problems. The stability of the reconstruction algorithm based on L1 was compared with L2. L2 failed in highly under-determined systems while L1 still gave stable and good results.

In addition, phantoms replicating the optical properties of biological tissues are designed and also experimental validations are performed. Two different type of phantoms were used:solid (made of homogenized milk and black ink) and solid (epoxy resin or silicon based). As a scattering compound for solid phantoms a TiO2 particles were used.

My PhD project was part of a bigger project were also a sufisticated optical fluorescent tomografic scanner was designed and constructed for small animal studies.

Figure: Reconstructed images of a) multiple embedded point sources in a diffusion medium (µa=0.1, µ's=10) by least squares ( b) and d) ) and linear optimization methods ( c) and e) ). Number of source-detector pairs in b) and c) was 1428 and in d) and e) was reduced to 228 for 1600 mesh elements.


Publications E. Janunts, T. Pöschinger, F. Eisa, A. Langenbucher, Modeling and experimental verification for a borad beam light transport in optical tomography, Z.Med.Phys,doi:10.1016/j.zemedi., 2010.06.005, article in press, (2010)

T. Pöschinger, E. Janunts, H. Brünner, A. Langenbucher, CCD based projectional imaging of fluorescent probes in tissue-like media: experimental setup and characterization, Z.Med.Phys, doi:10.1016/j.zemedi., 2009.05.002, article in press, (2010)

Z. Zhu, E. Janunts, T. Eppig, T. Sauer, A. Langenbucher, Iteratively re-weighted bi-cubic spline representation of corneal topography and its comparison to the standard methods, Z.Med.Phys, doi: 10.1016/j.zemedi., 2010.07.002, article in press, (2010)

E. Janunts, T. Pöschinger, H. Brünner, A. Langenbucher, Linear method of fluorescent source reconstruction in a diffusion medium, Z.Med.Phys., 18(3), 189-196, (2008)