The acronym NAT stands for Nonlinear and Adaptive Techniques in digital image processing, analysis and computer vision, an Esprit Basic Research Activity funded by the European Union. The partners in the activity are :
Nonlinear techniques in digital image processing and analysis have exhibited a tremendous growth in the past decade. Their use has been a successful alternative for the cases where the classical linear techniques fail. In this project, the application of nonlinear techniques in BW image processing and analysis, in image sequence processing as well as in color image processing are studied. Another important issue examined in the project is nonlinear digital filter adaptation. Novel adaptation algorithms are derived and novel adaptive nonlinear structures are investigated, coming from the merging of nonlinear filter families and neural networks. The aim of this project is to enhance the research efforts of the participating groups in the area of nonlinear and adaptive systems and techniques. The following nonlinear filter classes are studied: polynomial filters, order statistic filters and morphological filters. Adaptation methods for these filter classes are developed. Neural networks techniques are used for filter adaptation and digital image processing and analysis. All the above mentioned filters and their adaptive variations are used for :
In the case of nonlinear digital image analysis and computer vision, two approaches are investigated: mathematical morphology techniques and nonlinear position invariant transforms. Several of the above mentioned nonlinear techniques are prime candidates for parallel implementation. Parallel implementations are studied for all nonlinear processing and analysis schemes to be developed within this project.
So far, the project has produced novel contributions in all of the above mentioned areas. It has also addressed the need for a unifying theory for the various nonlinear filter classes by the interaction and exchange of research results in the various filter classes. Furthermore, the application of these nonlinear techniques in certain key-technology areas (e.g. color image processing, image sequence processing) has helped in the comparative study of these techniques and has produced a clear leading-edge technological output. Finally, the parallelisation of algorithms for nonlinear digital image processing and analysis will continue to contribute to the development of fast parallel unifying methods and architectures for this particular area.
Prof. Ioannis Pitas,
Dept. of Informatics,
Aristotle University of Thessaloniki,
P.O. Box 451, GR-54006 Thessaloniki, Greece
e-mail pitas@zeus.csd.auth.gr
Prof. Hans Burkhardt,
Technical University of Hamburg-Harburg,
P.O. Box 901052, 2100 Hamburg 90, Germany
e-mail burkhardt@tu-harburg.d400.de
Dr. Stephen Marshall,
Dept. Electronic & Electrical Eng,
University of Strathclyde,
204 George St, Glasgow, G1 1XW. UK
e-mail s.marshall@eee.strath.ac.uk
Dr. Moncef Gabbouj, Tampere University of Technology,
P.O. Box 553, 33101 Tampere, Finland
e-mail moncef@cs.tut.fi
Prof. Giovanni Sicuranza,
University of Trieste,
Via A. Valerio 10, 34127 Trieste, Italy
e-mail sicuranza@univ.trieste.it
NATTER
, the NAT newsletter, presents some of the advances made by the groups involved. So far, the following issues have been published, which can be viewed by clicking the highlighted words DOS or Unix, depending on the type of host your Web browser runs on :Copies of NAT documents can be obtained, through anonymous FTP at the following sites :
