Description

The laboratory is involved in basic research connected to biomedical signal and imaging data. The main research goals are summarized as following:

  • understanding and describing the physiological phenomena,
  • use of the computer in modeling and understanding of physiologic relationships,
  • locally and remotely monitoring physiologic events, graphically displayed anatomic details and physiologic functions, visualizing and representing biomedical signal and imaging data,
  • developing standardized databases to study physiologic mechanisms and to evaluate performance and robustness of recognition techniques,
  • characterizing data, and establishing the detection criteria and recognition techniques to automatize as much as possible the analysis of bioelectric patterns, examinations, procedures, and medical practice, in order to improve the quality and reliability of the examinations,
  • interpret the data and the results qualitatively and quantitatively,
  • develop performance measures and protocols to evaluate detection techniques,
  • develop biomedical information technologies and software.

The principal research topic currently underway is the development and evaluation of recognition algorithms for accurate detecting of transient ischaemic events in biomedical data using the LTST DB (Long-Term ST Database).

The second research topic is maintaining, updating and distribution of standardized international reference-annotated ECG database LTST DB. The database is result of a multinational research effort and contains 86 24-hour ambulatory recordings with a number of human-annotated transient ischaemic and non-ischaemic ST events. The database is intended to serve as a reference set in evaluating the performance of ST analyzers, and as a reference set to study physiologic mechanisms responsible for ischeamia. See:
http://www.physionet.org/physiobank/database/ltstdb/
and
http://www.physionet.org/challenge/2003/.

The next research topic is development of interactive graphic user interface editing tools (SEMIA – SEMI-Automatic) to visualize, display and annotate long-term electrocardiograms. SEMIA, version 3.0.1, to view diagnostic and morphology feature-vector time series, and to examine human annotations of the LTST DB is under GNU General Public Licence and is available from  http://www.physionet.org/physiobank/database/ltstdb/semia/.

The next research topic is the characterization of temporal patterns of transient ischaemic events and time-frequency representations of diagnostic parameters in ambulatory ECG signals. The goals are to differentiate physiologic mechanisms generating ischaemia and predicting impending ischaemia.

Another important contribution of the laboratory to the world community is interactive graphic tool EVAL_ST to evaluate performance and robustness of ischaemia analysers. The tool is under GNU General Public Licence and is available from  http://www.physionet.org/physiotools/eval_st/.

Another topic concerns the investigation and assessment of effective methods for monitoring patients affected by cardiovascular diseases, outside the specialized cardiac units, through computer analysis and the interpretation of non-invasive bio-signal data, with the ultimate goal of cardiac telemonitoring via the Internet.

Yet anoter very important achievement of the Laboratory for Biomedical Computer systems and imaging is developed Term-Preterm ElectroHysteroGram DataBase (TPEHG DB) in 2008. The aim of the Database is development of techniques to predict preterm delivery. The TPEHG DB is freely available on the following link:  http://lbcsi.fri.uni-lj.si/tpehgdb/. In 2010 we also post this database to Physionet.

The Laboratory supports a Web mirror site (http://physionet.fri.uni-lj.si) for a part of Europe to the PhysioNet Web site (http://www.physionet.org) which is located at the Massachusetts Institute of Technology in Cambridge, USA. Maintaining of the PhysioNet Web site is supported by the U.S. National Institutes of Health. Physionet offers free access via the Web to large collections of recorded physiologic signals and related open-source software.