301
Computational Vision and Medical Image Processing IV – João Tavares & Natal Jorge (eds)
© 2014 Taylor & Francis Group, London, ISBN 978-1-138-00081-0
Pattern recognition framework based on the best rank-(R
1
, R
2
, …, R
K
)
tensor approximation
B. Cyganek
AGH University of Science and Technology, Krakow, Poland
ABSTRACT: The paper presents a framework for pattern recognition in digital images based on the
best rank-(R
1
, R
2
, …, R
K
) decomposition of the prototype pattern tensors. The tensors are obtained from
the patterns defining the classes. In the case of a class with only a single prototype, its pattern tensor is
constructed from geometrically deformed versions of that pattern. Pattern recognition is accomplished
by testing a distance of the fea ...