Sim-Min-Hash: An efficient matching technique for linking large image collections - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2013

Sim-Min-Hash: An efficient matching technique for linking large image collections

Wan-Lei Zhao
Connectez-vous pour contacter l'auteur
Hervé Jégou
  • Function : Author
  • PersonId : 833473
Guillaume Gravier

Abstract

One of the most successful method to link all similar images within a large collection is min-Hash, which is a way to significantly speed-up the comparison of images when the underlying image representation is bag-of-words. However, the quantization step of min-Hash introduces important information loss. In this paper, we propose a generalization of min-Hash, called Sim-min-Hash, to compare sets of real-valued vectors. We demonstrate the effectiveness of our approach when combined with the Hamming embedding similarity. Experiments on large-scale popular benchmarks demonstrate that Sim-min-Hash is more accurate and faster than min-Hash for similar image search. Linking a collection of one million images described by 2 billion local descriptors is done in 7 minutes on a single core machine.
Fichier principal
Vignette du fichier
simminhash_hal.pdf (202.25 Ko) Télécharger le fichier
Vignette du fichier
SMH-timings.png (27.07 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Format : Figure, Image
Loading...

Dates and versions

hal-00839921 , version 1 (01-07-2013)
hal-00839921 , version 2 (01-07-2013)
hal-00839921 , version 3 (05-08-2013)

Identifiers

  • HAL Id : hal-00839921 , version 3

Cite

Wan-Lei Zhao, Hervé Jégou, Guillaume Gravier. Sim-Min-Hash: An efficient matching technique for linking large image collections. ACM Multimedia, Oct 2013, Barcelona, Spain. ⟨hal-00839921v3⟩
728 View
1920 Download

Share

Gmail Facebook X LinkedIn More