Segmentation using vector-attribute filters : methodology and application to dermatological imaging

Naegel, Benoît (School of Engineering, Architecture and Landscape (hepia), HES-SO // University of Applied Sciences Western Switzerland ; LORIA - Laboratoire Lorrain de Recherche en Informatique et ses Applications) ; Passat, Nicolas (LSIIT - Laboratoire des Sciences de l'Image, de l'Informatique et de la Télédétection) ; Boch, Nicolas (Digital Imaging Unit) ; Kocher, Michael (School of Engineering, Architecture and Landscape (hepia), HES-SO // University of Applied Sciences Western Switzerland)

Attribute-based filters can be involved in analysis and processing of images by considering attributes of various kinds (quantitative, qualitative, structural). Despite their potential usefulness, they are quite infrequently considered in the development of real applications. A cause of this underuse is probably the difficulty to determine correct parameters for non-scalar attributes in a fast and efficient fashion. This paper proposes a general definition of vector-attribute filters for grey-level images and describes some solutions to perform detection tasks using vector-attributes and parameters determined from a learning set. Based on these elements, an interactive segmentation method for dermatological application has been developed.


Keywords:
Conference Type:
full paper
Faculty:
Ingénierie et Architecture
School:
HEPIA - Genève
Institute:
inSTI - Institut des Sciences et Technologies industrielles
Publisher:
Rio de Janeiro, Brazil, 10-13 October 2007
Date:
2007-10
Rio de Janeiro, Brazil
10-13 October 2007
Pagination:
12 p.
Published in:
Proceedings of International Symposium on Mathematical Morphology (ISMM), 10-13 October 2007, Rio de Janeiro, Brazil
Numeration (vol. no.):
pp. 239-250
Appears in Collection:



 Record created 2020-05-01, last modified 2020-05-29

Fulltext:
Download fulltext
PDF

Rate this document:

Rate this document:
1
2
3
 
(Not yet reviewed)