On the burstiness of visual elements - Archive ouverte HAL Access content directly
Conference Papers Year : 2009

On the burstiness of visual elements


Burstiness, a phenomenon initially observed in text retrieval, is the property that a given visual element appears more times in an image than a statistically independent model would predict. In the context of image search, burstiness corrupts the visual similarity measure, i.e., the scores used to rank the images. In this paper, we propose a strategy to handle visual bursts for bag-of-features based image search systems. Experimental results on three reference datasets show that our method significantly and consistently outperforms the state of the art.
Vignette du fichier
burst.jpg (145.99 Ko) Télécharger le fichier Fichier principal
Vignette du fichier
jegou_burstiness.pdf (2.63 Mo) Télécharger le fichier
jegou_burstiness_poster.pdf (26.63 Mo) Télécharger le fichier
Format : Figure, Image
Origin : Files produced by the author(s)
Format : Other

Dates and versions

inria-00394211 , version 1 (23-02-2011)
inria-00394211 , version 2 (06-06-2011)
inria-00394211 , version 3 (07-06-2011)



Hervé Jégou, Matthijs Douze, Cordelia Schmid. On the burstiness of visual elements. CVPR 2009 - IEEE Conference on Computer Vision and Pattern Recognition, Jun 2009, Miami, United States. pp.1169-1176, ⟨10.1109/CVPRW.2009.5206609⟩. ⟨inria-00394211v3⟩
459 View
1028 Download



Gmail Facebook Twitter LinkedIn More