ML Anthology

On the Burstiness of Visual Elements

Hervé Jégou, Matthijs Douze, Cordelia Schmid
CVPR 2009 pp. 1169-1176
doi:10.1109/CVPR.2009.5206609 /cvpr/2009/jegou2009cvpr-burstiness/

Abstract

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.

CVPR Semantic Scholar

Cite

Text

Jégou et al. "On the Burstiness of Visual Elements." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2009. doi:10.1109/CVPR.2009.5206609

Markdown

[Jégou et al. "On the Burstiness of Visual Elements." IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2009.](https://mlanthology.org/cvpr/2009/jegou2009cvpr-burstiness/) doi:10.1109/CVPR.2009.5206609

BibTeX

@inproceedings{jegou2009cvpr-burstiness,
  title     = {{On the Burstiness of Visual Elements}},
  author    = {Jégou, Hervé and Douze, Matthijs and Schmid, Cordelia},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2009},
  pages     = {1169-1176},
  doi       = {10.1109/CVPR.2009.5206609},
  url       = {https://mlanthology.org/cvpr/2009/jegou2009cvpr-burstiness/}
}