Query-Adaptative Locality Sensitive Hashing
Abstract
It is well known that high-dimensional nearest-neighbor retrieval is very expensive. Many signal processing methods suffer from this computing cost. Dramatic performance gains can be obtained by using approximate search, such as the popular Locality-Sensitive Hashing. This paper improves LSH by performing an on-line selection of the most appropriate hash functions from a pool of functions. An additional improvement originates from the use of $E_8$ lattices for geometric hashing instead of one-dimensional random projections. A performance study based on state-of-the-art high-dimensional descriptors computed on real images shows that our improvements to LSH greatly reduce the search complexity for a given level of accuracy.
Fichier principal
qalsh.pdf (119.68 Ko)
Télécharger le fichier
qalsh.png (21.49 Ko)
Télécharger le fichier
icassp_qalsh_talk.pdf (823.16 Ko)
Télécharger le fichier
Origin : Files produced by the author(s)
Format : Figure, Image
Origin : Files produced by the author(s)