Projects

Details
Category: Science and education
Published on Wednesday, 23 May 2012 16:22
Written by Super User
Hits: 14359

Scientific research projects

  1. Systems and Algorithms for Signal and Image Processing, Nbr. 036028, 2002 -
  2. Digital speech processing in contemporary information technologies, Nbr. 036054, 2002 -
  3. Analog and digital system design based on optimization procedures, Br. 036029, 2002 -
  4. Systems and Algorithms for Signal and Image Processing, Nbr. 036-024, 1996-2001
    (Project leader: Prof. emer. Hrvoje Babić, PhD)
  5. Optimization of antialiasing filters, Nbr. 036-124, 1998 - 2001
    (Project leader: Mladen Vučić, PhD)
  6. Vector quantization of speech signal, Nbr. 036-224 1998 - 2001
    (Project leader: Davor Petrinović, PhD)
  7. Algorithms and systems for Signal Processing, Nbr. 2-07-265, 1993 - 1996
    (Project leader: Prof. emer. Hrvoje Babić, PhD)
  8. Systems and Algorithms for Signal Processing, Nbr. 2-07-364, 1993 - 1996
    (Project leader: Prof. emer. Hrvoje Babić, PhD)


Technological projects

  1. Implementation of Processor Cores on Programmable Gate Arrays, 2002 -
    (Project leader: Mladen Vučić, PhD)


Information technology projects

  1. Digital textbook for *NIX system administrators, 2003 (Project leader: Karmela Aleksić-Maslać, Ms)
  2. Design and Implementation of low-cost teleclassrooms, 2002 (Project leader: Damir Seršić, PhD)
  3. Electronic publications and distance learning materials for course in Signal and Systems, 2002 (Project leader: Prof. emer. Hrvoje Babić, PhD)
  4. Electronic publications and distance learning materials for course in Stochastic Processes in Systems, 2002 (Project leader: Prof. Sven Lončarić, PhD)
  5. Papyrus - Publishing system for electronic scientific journals, 2002 (Project leader: Prof. Sven Lončarić, PhD)
  6. Virtual polyclinic: web-based system for medical consultation, 2001 (Project leader: Prof. Sven Lončarić, PhD)

CEEPUS projects

  1. 3D Segmentation of Abdominal Aortic Aneurysm, 2000- (Project leader: Prof. Sven Lončarić, PhD)

 


Systems and Algorithms for Signal and Image Processing, Nbr. 036028, 2002-

Project leader: This email address is being protected from spambots. You need JavaScript enabled to view it.

Research within this project will be concentrated on the five research topics hence the five project sections. The first section is focused on subband signal decomposition, the second one is concentrated on algorithms for filter design, while the third is related to image processing as a special case of a two dimensional signal. The last two sections cover the field of computer system security and distance learning. In section dealing with subband signal decomposition new methods of adaptive linear and non-linear subband decomposition for multidimensional signal processing will be proposed. The aspects of implementation on general-purpose computers will be explored, with respect to applications in communications. In the section focused on filter design, new design methods will be developed. The methods will be based on the requirements that arise from emerging technologies such as communication and multimedia. In the image processing section new methods for image segmentation will be developed, that are based on clustering algorithms, mathematical morphology methods, fuzzy set concepts, genetic algorithms, neural networks, expert systems and deformable models. New robust methods that require minimal amount of human intervention will be proposed. The developed methods will be tested and optimized on real images and applied in medical diagnostics and production process automation. Contemporary computers, networks and information systems represent very complex electronic systems generating huge amount of signals. A special type of signals is computer and network logs, and their number and size is growing exponentially. They record network and user activities and contain all data about security incidents. However, because of their size, it is impossible to analyze all logs manually to recognize incidents. Since logs are a kind of multiparameter signal, i.e. a signal that is a sum of a number of various signals, we are dealing with the problem of extracting a known signal from the noise, or pattern recognition. It is expected that by applying know algorithms and techniques of digital signal processing, a significant number of incidents could be recognized, thus improving the security of information systems significantly. In section dealing with distance learning and multimedia, new educational methods will be developed using information and telecommunication technologies. Synchronous and asynchronous educational models will be thoroughly investigated.

(Published papers)

 

Digital speech processing in contemporary information technologies, Br. 036054, 2002 -

Project leader: This email address is being protected from spambots. You need JavaScript enabled to view it.

Digital processing of speech signals represents a foundation of many contemporary information technologies present in our everyday life. Some of these technologies include digital mobile telephony, systems for automatic speech and speaker recognition, speech synthesis systems, man-machine interfaces, voice over IP protocols, teleconferencing systems, voice mail, digital voice recorders etc. Methods for suitable parametric modeling of speech as well as quantization of model parameters represent a very important part of any speech coding system. Research within the scope of this project will therefore be oriented towards development of new speech modeling and speech coding methods and algorithms that can be applied in any of the above mentioned technologies. The emphasis will be laid upon development of efficient algorithms that have low algorithmic complexity while still insuring high quality reconstructed speech and low transmission rate. This is important because most of the systems that incorporate such algorithms are intended for real-time operation and therefore coding itself as a small part of the overall application should incur minimal processing load. Coding algorithms as well as speech and speaker recognition algorithms are usually developed for a specific language and their design is based on training speech databases in that language. It should be stressed that Croatian database will be created for the purpose of this project and it will be reusable in future related projects and research. Apart from developing new algorithms, project scope partially includes the survey of the available speech enabled technologies with their application aimed at advanced man-machine interfaces, modern multimedia systems for learning and teleconferencing systems. Since most of such applications involve the use of Internet, data privacy and protection represent important tasks that will also be addressed as part of the project. The investigations in this area will cover the aspects of computer system safety as well as the protection of information sent over computer networks. As a conclusion, the expected results of the project include intensive research of new efficient algorithms and methods for modeling and coding of speech, implementation of developed methods on several software and hardware platforms, development of new advanced man-machine communication systems and methods for safe and secure voice communication.

(Published papers)

 

Analog and digital system design based on optimization procedures, Br. 036029, 2002 -

Project leader: This email address is being protected from spambots. You need JavaScript enabled to view it.

In recent years, rapid increase of computing power has caused intensive work on the development and application of optimization methods. Based on the optimization procedures many new methods have arisen, covering the overall field of analog and digital system design. Research within the scope of our previous projects has shown that optimization methods can be successfully applied in the design of optimal analog and recursive digital filters as well as two-channel and M-channel filter banks. The goal of the proposed project is to improve the existing and to develop new optimization based methods for analog and digital system design. The research will start from the hypothesis that integral criteria for optimum filter design which are given in the time domain can successfully replace common criteria given in the frequency domain. New criteria will be synthesized together with the appropriate goal functions. Based on the new criteria, new methods for the design of optimal systems will be formed. The methods will be suitable for the design of analog and digital filters and correctors, filter banks and systems for wavelet signal analysis. New methods will contribute to the signal processing theory and applications. The contribution will be noticeable in new structures and methods that engineers can use in their work. From theoretical point of view, the results obtained by using new methods will give an additional insight in the behavior of the systems. Furthermore, the theoretical analysis will combine the existing and the new knowledge, possibly generalizing some actual problems in the field of analog and digital system design. Signal processing is one of few areas in which the results of a basic research can be directly applied. For example, in digital signal processing, algorithms supported by their signal flow schemes should only be transformed to the appropriate software and downloaded to usually standard hardware environment. Therefore, the development of an algorithm, i.e. the theoretical work, might be the most difficult and the most important part of the software or even the overall system design. The knowledge in signal processing can be offered on the global market, regardless of the actual situation in the Crotaian high-technology sector. The topics covered by the project can be applied in communications and multimedia.

(Published papers)

FaLang translation system by Faboba