10th VIPS Advanced school on Computer Vision and Pattern Recognition

Verona 23-26 Semptember 2013

Dissimilarity-based Representation for Pattern Recognition

image

Prof. Robert Duin

Delft University of Technology

home: http://rduin.nl

The dissimilarity representation allows the use of non-Euclidean and asymmetric dissimilarity measures. It is thereby a powerful approach for users to apply different ways to express their domain knowledge than by features or by normalizing and sampling the objects to be recognized. The lectures cover discussions of the principles, explanations of the tools and presentations of some applications.
The course will cover the dissimilarity-based paradigm for pattern recognition. It offers an alternative for the traditional feature based approaches. Based on user defined dissimilarity measures between raw object representations like images, time signals and spectra vector representations are defined enabling the use of established tools for dimension reduction and classification.The course will be composed by theoretical as well practical lessons, using Matlab and the PRTOOLS library (see www.prtools.org and www.37steps.com). 

For any other information, please send an email to manuele.bicego@univr.it or umberto.castellani@univr.it;