Paper Info
Title
Image classification with a frequency-based information retrieval scheme for ImageCLEFmed 2006
Abstract
This article describes the participation of the University and Hospitals of Geneva at the ImageCLEF 2006 image classification tasks (medical and non-medical). The techniques applied are based on classical tf/idf weightings of visual features as used in the GIFT (GNU Image Finding Tool). Based on the training data, features appearing in images of the same class are weighted higher than features appearing across classes. These feature weights are added to the classical weights. Several weightings and learning approaches are applied as well as quantisations of the features space with respect to grey levels. A surprisingly small number of grey levels leads to best results. Learning can improve the results only slightly and does not obtain as good results as classical image classification approaches. A combination of several classifiers leads to best final results, showing that the schemes have independent results.
Year
DOI
Venue
2006
10.1007/978-3-540-74999-8_78
CLEF
Keywords
Field
DocType
grey level,classical tf,image classification task,classical image classification approach,frequency-based information retrieval scheme,classical weight,features space,best result,gnu image,idf weightings,best final result,classification,information retrieval,image classification,feature space,image retrieval
Small number,Training set,Feature detection (computer vision),Information retrieval,Pattern recognition,Feature (computer vision),Computer science,Image retrieval,Artificial intelligence,Natural language processing,Contextual image classification,Visual Word
Conference
Volume
ISSN
ISBN
4730
0302-9743
3-540-74998-5
Citations 
PageRank 
References 
0
0.34
7
Authors
3
Name
Order
Citations
PageRank
Henning Müller12538218.89
Tobias Gass214311.91
Antoine Geissbuhler381549.75