Applications of PDE-Based Image Processing in Fluorescence Microscopy

Investor logo

Warning

This publication doesn't include Faculty of Economics and Administration. It includes Faculty of Informatics. Official publication website can be found on muni.cz.
Authors

HUBENÝ Jan

Year of publication 2008
MU Faculty or unit

Faculty of Informatics

Citation
Description With the recent development of high-resolution microscopes and fluorescent probes together with tremendous development of information technologies, the cell biology has entered a new era. In these days, the biologists are able to observe sub-cellular components like chromosomes or even individual genes and proteins, the observations can be made in two or three dimensions and even dynamic processes in the living cell can be investigated. However, the research in this field depends more and more on imaging and computer vision methods and computer vision specialists, because such detailed observations and investigations produce huge image data sets which cannot be analyzed by hand. A large number of computer vision methods have been developed in past two decades to support research in this field. However, the majority of successful methods is based on rather elementary image processing techniques. The image processing and computer vision have gone through a similar tremendous development as the cell biology in past two decades. The specialists in this field have developed a large amount of sophisticated mathematically well-founded image processing methods. The PDE-based (Partial Differential Equation) image processing techniques belong to such methods. However, these new methods are still not very often used in biological imaging, either for their complexity or for their computational time demands. In this thesis, we present several applications of PDE-based image processing methods in fluorescence microscopy. We first provide a brief background on the fluorescence microscopy and its usage in the cell biology, then we review a selection of PDE-based image processing methods. As an own contribution, a segmentation algorithm based on two recent approximations of Chan-Vese active contour model is presented and tested. Further, two methods for the segmentation of interphase chromosome territories are designed, presented and tested. Finally, we present the application of state-of-the-art variational optic flow methods for motion estimation in time-lapse sequences from live-cell studies.
Related projects:

You are running an old browser version. We recommend updating your browser to its latest version.