Special PhD positions at the Department of Visual Computing


Department of Visual Computing, Faculty of Informatics, Masaryk University, announces an open call for one PhD position starting from the Autumn 2019 term, with applications in the following areas:

The deadline for application is May 15, 2019.

General information

The student is expected to have (or be about to finish) a Master level education in Computer Science, Electrical Engineering, Biomedical Engineering, or related areas, with demonstrated overview in the field. Good knowledge of English language is expected as well as willingness to spend 3-6 months in a collaborating group abroad during the PhD studies; prior knowledge of Czech is not necessary.


The applications will be evaluated by the department committee, whose members will choose the best applicant. The announced PhD position is funded with an extra department stipend of 10 000 CZK per month (summing to at least 28 700 CZK with the standard faculty stipend). The stipend is granted to the successful applicant for the first 2 years, with an expected renewal (after an evaluation) for another 2 years. The total length of study is 4 years.

Application procedure

Applicants are advised to contact directly their perspective supervisor (as listed below) for more specific details, well ahead of the deadline. The final applications consisting of CV (including education, degrees and dates, publications/scientific presentations, skills/experiences in programming languages, project work, academic awards, etc.), motivation letter explaining why you apply specifically for this project and why you are the perfect candidate, transcript of the grades from the Master’s and Bachelor's degree, two references (written or just two contact names), and possibly other relevant documents supporting the candidate's excellence should be sent to the Head of the Department

  • Assoc.Prof. Petr Matula, before the respective deadline for application (see above)

The candidates are still obliged to pass the standard admission procedure for doctoral study. The stipend can be awarded only after successfully completing the standard admission procedure.

Topic: Cell Behaviour Studies using Machine Learning

Supervisor: prof. RNDr. Michal Kozubek, Ph.D.
Area: Biomedical image processing

Understanding the cell in its spatiotemporal context is the key to unraveling many of the still unknown mechanisms of life and disease, hence there are ongoing efforts to integrate all the vast and diverse information about cells to create a credible model of cell morphology and behavior [1]. While cell morphology has been studied rather thoroughly for decades, there is still lack of information on cell behavior under various conditions. Therefore, the goal of the PhD student will be to contribute to the learning and understanding of cell behavior, especially by applying machine learning methods to the analysis of a large collection of videos from optical microscopes. Both publicly available datasets and new videos of cells will be analyzed. Know-how on cell tracking will be provided so that the student will be able to concentrate on the behavior analysis. At the beginning of the work it will be necessary to define a suitable collection of behavior descriptors (temporal features) – there are not many of them in contrast to morphology descriptors (spatial features). Afterwards, the cell behavior will be studied using these descriptors. And because cells are sociable entities, it will be necessary to study not only the behavior of single cells but also groups of cells and interactions between cells. Eventually, the computer should be able to predict subsequent behavior of cells (key temporal features) based on their previous behavior, i.e. predict “the rest of the story” for an unfinished video (e.g., that the cells will die or form some structure). This would have an immense impact, e.g. for quality control in stem cell research (to check that stem cells behave in a correct way).

[1] Ortiz-de-Solórzano C, Muñoz-Barrutia A, Meijering E, Kozubek M. Toward a Morphodynamic Model of the Cell. IEEE Signal Processing Magazine, New York: IEEE, 2015, vol. 32, No 1, p. 20-29. ISSN 1053-5888. 2015. doi:10.1109/MSP.2014.2358263.

Topic: Brain Computer Interfaces for Virtual and Augmented Reality

Supervisor: assoc. prof. Fotis Liarokapis
Area: Human computer interaction

BCIs often make use electroencephalography (EEG) as neuro-feedback in a number of application domains. In a BCI system, brain modulation patterns can be extracted and analyzed in order to determine the mental state of the user. These states can be translated with the help of signal processing algorithms and machine learning into a control signal that could act as an input for computers or external devices. Non-invasive BCIs are getting a lot of attention as alternative human-computer interaction devices for games and virtual environments. Non-invasive BCIs operate by recording the brain activity from the scalp with Electroencephalography (EEG) sensors attached to the head on an electrode cap or headset without being surgically implanted. However, they still have a number of problems and they cannot function as accurately as other natural user interfaces (NUIs) and traditional input devices such as the standard keyboard and mouse. The aim of this research topic is to investigate non-invasive BCI technologies for human-computer interaction using visual stimulus generated either by virtual or augmented reality.

Topic: Cell Tracking in Microscopy

Supervisor: assoc. prof. Pavel Matula
Area:Biomedical image processing

The latest results on comparison of cell-tracking algorithms [1] indicated that the state-of-the-art algorithms are still, despite the very good results for some datasets, in general far from the demanded outcome. Especially, the development is necessary for scenarios with low signal-to-noise ratio or low contract ratio or for tracking cells with more complex shapes or textures. Large 3D data sets, such as those of developing embryos, present additional challenges. Not only do such videos show very high cell densities in later frames, the size of the image data itself causes very long runtimes. The aim of this research topic is to contribute to the solution of these chalenging problems.

[1] Ulman et. al. An objective comparison of cell-tracking algorithms, Nature Methods 2017

Supervisor: doc. RNDr. Petr Sojka, Ph.D.
Area: Knowledge representation

The key for successful understanding and retrieval of published knowledge stored in digital libraries are semantic representations that capture latent narrative and structural qualities of fulltexts. Accurate representation with scalable similarity search would allow to achieve precision that is hard to achieve with today's topic based approaches with vector based representations. The aim of the thesis is to build both evaluation methodology and respective system capable of retrieval scientific papers of arXiv type taking into account both semantic, narrative and structural qualities ("semantic sketches" or "trains of thoughts") of indexed fulltexts.

Publications relevant to the given topic can be found on the pages of MIR group.