Image analysis in general

Let us start by omitting the adjective biomedical and talk about what the image analysis is and what it involves.

Imagine that we have spent summer holidays by the sea. We might have captured lots of images using a digital camera and later on we've uploaded them into a computer. A task, for instance, for image analysis could be to count how many ships are there in the given image of the sea. To estimate the average length of ships would be perhaps even more interesting task for image analysis. In this case, one would have to consider the distance of every ship to the photographer owing to the presence of the perspective projection. In addition, one would have not only to detect every single ship in the image reliably but also he or she must find ship's outline very precisely.

In general, we may say that during image analysis we are trying to extract certain information from an image by assigning a certain interpretation to image data. The key is to find a good interpretation assignment, which often requires to combine knowledge from several fields, not necessarily always from the computer science. It is true that even a simple task usually requires nontrivial solution. In the example given above, we need to be able to find a ship in the image, that is to describe a general ship in terms of digital image data (the assignment), that is to characterize it or, eventually, to define what makes it different in its region. There is even more, we also need to know the transformation from image length unit to a real world one to finally yield ships' lengths.


Biomedical image analysis at the Faculty of informatics

It is apparent that image analysis often utilizes a knowledge from some other field or fields. In the Centre for Biomedical Image Analysis (abbreviated to CBIA) are these additional fields the molecular biology and genetics.

Example of a real task: to track and visualize movements of stained cellular structures.
an image of a cell and a visualization of movements of cell structures of interest image of a cellmovement visualization

It is very difficult to be an expert in multiple fields simultaneously. Hence, the research is often interdisciplinary. The same holds for CBIA: computer scientists are seeking reliable, accurate and automatic solutions for image processing and analysis to aid the research in biology while biologists are delivering their expertise and experience. Both fields are sort of helping each other.

Research projects

  • We suppress noise and aberrations of the microscope optical system
  • We track positions, movements and shape changes of cellular structures over time
  • We analyse gene activities and seek mutual dependences and relation to deleterious human diseases

Development projects

Applications

  • FISH - the first software at FI to automatically control a microscope and analyse images
  • Acquiarium - an improved modern software for truly automatic image acquisition
  • JABS - analysis of microarray images
  • CytoPacq - a tool to generate artificial biomedical image data and simulate its acquisition using optical microscopy

Libraries

  • i3dcore - a template multi-platform library written in C++ designed for reading, representing and writing 3D images
  • i3dalgo - a template multi-platform library written in C++ designed for 3D image processing and analysis
  • OpticalFlow - a library implementing 3D extensions of selected variational methods for computing the optical flow
  • v3dwins - a library to display 3D images using maximum and orthogonal projections

More can be found at the CBIA web page.

The team

CBIA logo

Grant projects in progress

  • Coordination of the Research project (No. 2B0652) "Determination of markers, screening and early diagnostics of cancer diseases using highly automated processing of multidimensional biomedical images"
  • Participation on the Research centre (No. LC535) "Chromosome dynamics and structure during the cell cycle"
  • Participation on two long-term research plans of Masaryk University

How to join us

The most easiest way to cooperate with CBIA is to register for bachelor, master and/or dissertation thesis. Clearly, theses push students to gradually take more specialized courses. Most of these include practical sessions in a computer lab to practise new skills. Short-term projects in the field of image processing are offered in PV162 and PV163 in both spring and autumn semesters.

A unique study field "Image processing" has been opened at the Faculty of informatics recently to answer the increasing demand for specialists in this field. A graduate is able to design and lead a development of software systems for processing and analysis of image data in the research (for instance, a molecular-biological research with the aid of microscope imaging techniques), in the medicine (for instance, processing and analysis of ultrasonic, MRI or CT images) as well as in the industry (for instance, recognition of fingerprints or retina patterns, security or traffic video surveillance and so on).

More information about courses offered can be found on the CBIA web.


Contacts

  • CBIA web presentation: http://cbia.fi.muni.cz/
  • List of currently offered theses topics: MU Information system
  • List of currently solved theses: CBIA web pages
  • Information about current short-term projects: the first lesson of PV162 and PV163
  • Email for questions, suggestions: cbia-info(atsign)fi(dot)muni(dot)cz/