Learning to Rank Histograms for Object Retrieval

Danfeng Qin, Yuhua Chen, Matthieu Guillaumin and Luc Van Gool

In Proceedings British Machine Vision Conference 2014


Most state-of-the-art object retrieval systems rely on ad-hoc similarities between histograms of quantised local descriptors to find, in their databases, all the images relevant to an image query. In this work, our goal is to replace those similarities with ones that are specifically trained to maximize the retrieval accuracy. We propose to use a simple and very general linear model whose weights directly represent the similarity values. We devise a variant of rank-SVM to learn those weights automatically from training data with fast convergence and we propose techniques to limit the weights to a tractable number. Importantly, the flexibility of our model allows us to seamlessly incorporate well-known image retrieval schemes such as burstiness, negative evidence and idf weighting, and still exploit inverted files for efficiency in the large-scale setting. In our experiments, we show that our approach consistently and significantly outperforms the similarities used in several state-of-the-art systems on 4 standard benchmark datasets. In particular, on the Oxford105k dataset, our method outperforms the closest competitor by 6%.


Poster Session


Extended Abstract (PDF, 1 page, 93K)
Paper (PDF, 12 pages, 792K)
Bibtex File


Danfeng Qin, Yuhua Chen, Matthieu Guillaumin, and Luc Van Gool. Learning to Rank Histograms for Object Retrieval. Proceedings of the British Machine Vision Conference. BMVA Press, September 2014.


	title = {Learning to Rank Histograms for Object Retrieval},
	author = {Qin, Danfeng and Chen, Yuhua and Guillaumin, Matthieu and Van Gool, Luc},
	year = {2014},
	booktitle = {Proceedings of the British Machine Vision Conference},
	publisher = {BMVA Press},
	editors = {Valstar, Michel and French, Andrew and Pridmore, Tony}
	doi = { http://dx.doi.org/10.5244/C.28.43 }