org.openimaj.lsh.functions

Class IntHammingFactory

java.lang.Object

org.openimaj.lsh.functions.RandomisedHashFunctionFactory<int[]>

org.openimaj.lsh.functions.IntHashFunctionFactory

org.openimaj.lsh.functions.IntHammingFactory

All Implemented Interfaces:

HashFunctionFactory<int[]>

@Reference(type=Inproceedings,
           author={"Indyk, Piotr","Motwani, Rajeev"},
           title="Approximate nearest neighbors: towards removing the curse of dimensionality",
           year="1998",
           booktitle="Proceedings of the thirtieth annual ACM symposium on Theory of computing",
           pages={"604","","613"},
           url="http://doi.acm.org/10.1145/276698.276876",
           publisher="ACM",
           series="STOC \'98")
public class IntHammingFactory
extends IntHashFunctionFactory

A hash function factory for producing hash functions that approximate the Hamming distance.

Author:

Jonathon Hare (jsh2@ecs.soton.ac.uk)

Field Summary

Fields inherited from class org.openimaj.lsh.functions.RandomisedHashFunctionFactory

ndims, rng

Constructor Summary

Constructors

Constructor and Description
IntHammingFactory(int ndims, cern.jet.random.engine.MersenneTwister rng, int bitsPerDim) Construct a new factory using the given parameters.

Method Summary

Modifier and Type	Method and Description
org.openimaj.lsh.functions.IntHammingFactory.Function	create() Construct a new HashFunction.
IntFVComparison	fvDistanceFunction()

Methods inherited from class org.openimaj.lsh.functions.IntHashFunctionFactory

distanceFunction

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

IntHammingFactory

public IntHammingFactory(int ndims,
                         cern.jet.random.engine.MersenneTwister rng,
                         int bitsPerDim)

Construct a new factory using the given parameters.

Parameters:

ndims - The number of dimensions (i.e. length of the vector being hashed)

rng - A random number generator

bitsPerDim - The number of bits per dimension. If the data is packed, then this will be greater than zero, and internally a single bit will be sampled for computing the hash. If zero, then it is assumed that every element of the vector being hashed is either a zero or one.

Method Detail

create

public org.openimaj.lsh.functions.IntHammingFactory.Function create()

Description copied from interface: HashFunctionFactory

Construct a new HashFunction.

Returns:

the new HashFunction

fvDistanceFunction

public IntFVComparison fvDistanceFunction()

Specified by:

fvDistanceFunction in class IntHashFunctionFactory