HPC and big data processing

group leader

prof. Sven Gotovac, PhD

associates

prof. Stipo Čelar, PhD
prof. Ivan Marinović, PhD
prof. Linda Vicković, PhD
assoc. prof. Vladimir Pleština, PhD
assist. prof. Dunja Božić-Štulić

Research topics

  1. Modelling of large distributed computer systems
  2. Processing of images from a drone: machine learning + convolution neural network
  3. Computer Vision in medical imaging
  4. Machine learning on big data
  5. Optimal computing architecture for real time processing
  6. Algorithm optimisation for dedicated computer architecture

Description of laboratory and equipment

  • Server computers, workstation with graphics card and PhI card
  • A server with four graphics cards
  • CERN and SRCE computer resources will be used for longer-term calculations

Contacts with academic and other institutions

  • European Organization for Nuclear Research (CERN), Geneva, Switzerland
  • University of Zagreb, Faculty of Science (PMFST), Croatia
  • University of Mostar, Faculty of Science and Education (FPMOZ), Bosnia and Herzegovina
  • University of Mostar, Faculty of Mechanical Engineering, Computing and Electrical Engineering (FSRE), Bosnia and Herzegovina
  • Hrvatski telekom Split, Croatia
  • HT ERONET Mostar, Bosnia and Herzegovina
  • Hrvatska vatrogasna zajednica
  • University Computing Centre (SRCE), Zagreb, Croatia
  • Vienna University of Technology, Austria
  • Norwegian University of Science and Technology (NTNU), Trondheim, Norway
  • Javna ustanova Vatrogasni centar za edukaciju i tehnološki razvoj Split – Vučevica “VA.C.E.T.RA.S”, Croatia
project title

High Performance Intelligent distributed system for big data processing (HPC-DSIOVP)

Description of research in a 5-year term

Research related to the parallelization of algorithms regarding simulations at CERN and modification of the code to run on AMD graphics cards.

In collaboration with the HVZ, research has been initiated on the online processing of images from unmanned aerial vehicles and their transmission via secure channels (4G or satellite) for offline processing. The use of modern machine learning techniques and convolutional neural networks to recognize people in the field and to cluster the field for the purpose of action planning. The models developed will be used in agriculture and ecology. The research is also being carried out as part of the IPA Interreg project SmartProtect, in which the FESB is involved.

The University Health Study is concerned with the automatic analysis of medical images, with a focus on the analysis of knee MRIs. PhD student Ana Pinjuh is working on this problem.

Development of a system for intelligent monitoring of traffic in ports and on the coast in cooperation with the Split Port Authority, Tehnet and Statim using the 4G/5G network, control of aircraft from the control center and transmission and processing of data on aircraft and the HPC center.