000146018 001__ 146018
000146018 005__ 20190101013055.0
000146018 037__ $$aREP_WORK
000146018 245__ $$aTagging and Retrieving Images with Co-Occurrence Models: from Corel to Flickr
000146018 269__ $$a2009
000146018 260__ $$bIdiap$$c2009
000146018 336__ $$aReports
000146018 520__ $$aThis paper presents two models for content-based automatic image annotation and retrieval in web image repositories, based on the co-occurrence of tags and visual features in the images. In particular, we show how additional measures can be taken to address the noisy and limited tagging problems, in datasets such as Flickr, to improve performance. An image is represented as a bag of visual terms computed using edge and color information. The first model begins with a naive Bayes approach and then improves upon it by using image pairs as single documents to significantly reduce the noise and increase annotation performance. The second method models the visual features and tags as a graph, and uses query expansion techniques to improve the retrieval performance. We evaluate our methods on the commonly used 150 concept Corel dataset, and a much harder 2000 concept Flickr dataset.
000146018 700__ $$aGarg, Nikhil
000146018 700__ $$0241066$$aGatica-Perez, Daniel$$g171600
000146018 8564_ $$uhttp://publications.idiap.ch/downloads/reports/2009/Garg_Idiap-RR-21-2009.pdf$$zURL
000146018 8564_ $$s2063303$$uhttps://infoscience.epfl.ch/record/146018/files/Garg_Idiap-RR-21-2009.pdf$$zn/a
000146018 909C0 $$0252189$$pLIDIAP$$xU10381
000146018 909CO $$ooai:infoscience.tind.io:146018$$pSTI$$preport$$qGLOBAL_SET
000146018 937__ $$aLIDIAP-REPORT-2009-004
000146018 970__ $$aGarg_Idiap-RR-21-2009/LIDIAP
000146018 973__ $$aEPFL$$sPUBLISHED
000146018 980__ $$aREPORT