Optimal LSH

Slaney, Malcolm Optimal LSH.

Summary

This package provides code to implement locality-sensitive hashing (LSH) in an optimum fashion.

There are two pieces. A Python library that implements LSH and a Matlab routine that calculates the optimum parameters for LSH.

The LSH implementation is based on a tutorial published by IEEE Malcolm Slaney, Michael Casey, "Locality-Sensitive Hashing for Finding Nearest Neighbors [Lecture Notes]," Signal Processing Magazine, IEEE , vol.25, no.2, pp.128-131, March 2008, doi: 10.1109/MSP.2007.914237
URL: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=4472264&isnumber=4472102 and also available at this URL http://www.slaney.org/malcolm/yahoo/Slaney2008-LSHTutorial.pdf

The optimization algorithm is based on this article, which is currently under review. Send email to malcolm@ieee.org for a preprint Malcolm Slaney, Yury Lifshits, Junfeng He, "Optimal Locality-Sensitive Hashing," Submitted to Proceedings of the IEEE, Special Issue on Web-Scale Multimedia, Summer 2012.

Information
Code Meta
Library
3:23
[img]
Github Repository
Optimal-LSH - Published Version

Download (70kB)
3:24
[img]
Preview
Figure 1 from Slaney & Casey 2008
Screenshot 2019-03-02 at 16.48.35.png

Download (1MB) | Preview
3:25
[img]
Preview
REF 2014 Impact Case Study
43819.pdf - Published Version

Download (206kB) | Preview
3:26
[img]
YahooArchive_Optimal-LSH.json

Download (6kB)
View Item