Query-Adaptative Locality Sensitive Hashing - Archive ouverte HAL Access content directly
Conference Papers Year : 2008

Hervé Jégou
• Function : Author
• PersonId : 833473
Laurent Amsaleg
Cordelia Schmid
• Function : Author
• PersonId : 831154
Patrick Gros

#### 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.

### Dates and versions

inria-00318614 , version 1 (15-03-2011)

### Identifiers

• HAL Id : inria-00318614 , version 1
• DOI :

### Cite

Hervé Jégou, Laurent Amsaleg, Cordelia Schmid, Patrick Gros. Query-Adaptative Locality Sensitive Hashing. ICASSP 2008 - IEEE International Conference on Acoustics, Speech, and Signal Processing, Mar 2008, Las Vegas, United States. pp.825-828, ⟨10.1109/ICASSP.2008.4517737⟩. ⟨inria-00318614⟩

### Export

BibTeX TEI Dublin Core DC Terms EndNote Datacite

331 View