Abstract | ||
---|---|---|
There is a growing need for efficient visual informationretrieval systems which take into account particularfeatures of images, in order to avoid nonsense results. Inthis paper we propose a new method for content-based imageretrieval, which can be divided into two main parts. 1)Automatic segmentation and extraction of shapes from imagesub-regions. 2) Ontological descriptions of shapescontained in images. Our method allows users to performsketch queries; these are simple drawings whichrepresent the main idea of something the user is expectingto retrieve from the results. Once the user introduces asketch, it is simplified using discrete curve evolution, then,it is turned into a tangent space representation. Similaritymeasures based on turning functions are used to recoversimilar images. An ontology reduces the proportion of non-senseresults since the ontology permits the recovery ofimages related semantically, even if they do not have similarshapes |
Year | DOI | Venue |
---|---|---|
2004 | 10.1109/ICECC.2004.1269588 | CONIELECOMP |
Keywords | Field | DocType |
main idea,visual information retrieval,content-based imageretrieval,efficient visual informationretrieval system,account particularfeatures,automatic segmentation,new method,discrete curve evolution,inthis paper,ontological shape-description,ontological description,main part,feature extraction,image segmentation,image retrieval | Ontology (information science),Similarity (geometry),Pattern recognition,Information retrieval,Computer science,Segmentation,Image retrieval,Feature extraction,Image segmentation,Artificial intelligence,Content-based image retrieval,Sketch | Conference |
ISSN | ISBN | Citations |
2474-9036 | 0-7695-2074-X | 2 |
PageRank | References | Authors |
0.54 | 4 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Alberto Chávez-Aragón | 1 | 15 | 4.61 |
Oleg Starostenko | 2 | 44 | 14.87 |