Results 1  10
of
39
Approximate Nearest Neighbors: Towards Removing the Curse of Dimensionality
, 1998
"... The nearest neighbor problem is the following: Given a set of n points P = fp 1 ; : : : ; png in some metric space X, preprocess P so as to efficiently answer queries which require finding the point in P closest to a query point q 2 X. We focus on the particularly interesting case of the ddimens ..."
Abstract

Cited by 1019 (40 self)
 Add to MetaCart
The nearest neighbor problem is the following: Given a set of n points P = fp 1 ; : : : ; png in some metric space X, preprocess P so as to efficiently answer queries which require finding the point in P closest to a query point q 2 X. We focus on the particularly interesting case of the ddimensional Euclidean space where X = ! d under some l p norm. Despite decades of effort, the current solutions are far from satisfactory; in fact, for large d, in theory or in practice, they provide little improvement over the bruteforce algorithm which compares the query point to each data point. Of late, there has been some interest in the approximate nearest neighbors problem, which is: Find a point p 2 P that is an fflapproximate nearest neighbor of the query q in that for all p 0 2 P , d(p; q) (1 + ffl)d(p 0 ; q). We present two algorithmic results for the approximate version that significantly improve the known bounds: (a) preprocessing cost polynomial in n and d, and a trul...
An Optimal Algorithm for Approximate Nearest Neighbor Searching in Fixed Dimensions
 ACMSIAM SYMPOSIUM ON DISCRETE ALGORITHMS
, 1994
"... Consider a set S of n data points in real ddimensional space, R d , where distances are measured using any Minkowski metric. In nearest neighbor searching we preprocess S into a data structure, so that given any query point q 2 R d , the closest point of S to q can be reported quickly. Given any po ..."
Abstract

Cited by 987 (32 self)
 Add to MetaCart
Consider a set S of n data points in real ddimensional space, R d , where distances are measured using any Minkowski metric. In nearest neighbor searching we preprocess S into a data structure, so that given any query point q 2 R d , the closest point of S to q can be reported quickly. Given any positive real ffl, a data point p is a (1 + ffl)approximate nearest neighbor of q if its distance from q is within a factor of (1 + ffl) of the distance to the true nearest neighbor. We show that it is possible to preprocess a set of n points in R d in O(dn log n) time and O(dn) space, so that given a query point q 2 R d , and ffl ? 0, a (1 + ffl)approximate nearest neighbor of q can be computed in O(c d;ffl log n) time, where c d;ffl d d1 + 6d=ffle d is a factor depending only on dimension and ffl. In general, we show that given an integer k 1, (1 + ffl)approximations to the k nearest neighbors of q can be computed in additional O(kd log n) time.
Similarity search in high dimensions via hashing
, 1999
"... The nearest or nearneighbor query problems arise in a large variety of database applications, usually in the context of similarity searching. Of late, there has been increasing interest in building search/index structures for performing similarity search over highdimensional data, e.g., image dat ..."
Abstract

Cited by 622 (13 self)
 Add to MetaCart
The nearest or nearneighbor query problems arise in a large variety of database applications, usually in the context of similarity searching. Of late, there has been increasing interest in building search/index structures for performing similarity search over highdimensional data, e.g., image databases, document collections, timeseries databases, and genome databases. Unfortunately, all known techniques for solving this problem fall prey to the \curse of dimensionality. &quot; That is, the data structures scale poorly with data dimensionality; in fact, if the number of dimensions exceeds 10 to 20, searching in kd trees and related structures involves the inspection of a large fraction of the database, thereby doing no better than bruteforce linear search. It has been suggested that since the selection of features and the choice of a distance metric in typical applications is rather heuristic, determining an approximate nearest neighbor should su ce for most practical purposes. In this paper, we examine a novel scheme for approximate similarity search based on hashing. The basic idea is to hash the points
When Is "Nearest Neighbor" Meaningful?
 In Int. Conf. on Database Theory
, 1999
"... . We explore the effect of dimensionality on the "nearest neighbor " problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance ..."
Abstract

Cited by 403 (1 self)
 Add to MetaCart
. We explore the effect of dimensionality on the "nearest neighbor " problem. We show that under a broad set of conditions (much broader than independent and identically distributed dimensions), as dimensionality increases, the distance to the nearest data point approaches the distance to the farthest data point. To provide a practical perspective, we present empirical results on both real and synthetic data sets that demonstrate that this effect can occur for as few as 1015 dimensions. These results should not be interpreted to mean that highdimensional indexing is never meaningful; we illustrate this point by identifying some highdimensional workloads for which this effect does not occur. However, our results do emphasize that the methodology used almost universally in the database literature to evaluate highdimensional indexing techniques is flawed, and should be modified. In particular, most such techniques proposed in the literature are not evaluated versus simple...
Similarity Indexing: Algorithms and Performance
 In Proceedings SPIE Storage and Retrieval for Image and Video Databases
, 1996
"... Efficient indexing support is essential to allow contentbased image and video databases using similaritybased retrieval to scale to large databases (tens of thousands up to millions of images). In this paper, we take an in depth look at this problem. One of the major difficulties in solving this pr ..."
Abstract

Cited by 127 (1 self)
 Add to MetaCart
(Show Context)
Efficient indexing support is essential to allow contentbased image and video databases using similaritybased retrieval to scale to large databases (tens of thousands up to millions of images). In this paper, we take an in depth look at this problem. One of the major difficulties in solving this problem is the high dimension (6100) of the feature vectors that are used to represent objects. We provide an overview of the work in computational geometry on this problem and highlight the results we found are most useful in practice, including the use of approximate nearest neighbor algorithms. We also present a variant of the optimized kd tree we call the VAM kd tree, and provide algorithms to create an optimized Rtree we call the VAMSplit Rtree. We found that the VAMSplit Rtree provided better overall performance than all competing structures we tested for main memory and secondary memory applications. We observed large improvements in performance relative to the R*tree and SStree in secondary memory applications, and modest improvements relative to optimized kd tree variants.Nearest Neighbor Search
A local search approximation algorithm for kmeans clustering
, 2004
"... In kmeans clustering we are given a set of n data points in ddimensional space ℜd and an integer k, and the problem is to determine a set of k points in ℜd, called centers, to minimize the mean squared distance from each data point to its nearest center. No exact polynomialtime algorithms are kno ..."
Abstract

Cited by 105 (1 self)
 Add to MetaCart
(Show Context)
In kmeans clustering we are given a set of n data points in ddimensional space ℜd and an integer k, and the problem is to determine a set of k points in ℜd, called centers, to minimize the mean squared distance from each data point to its nearest center. No exact polynomialtime algorithms are known for this problem. Although asymptotically efficient approximation algorithms exist, these algorithms are not practical due to the very high constant factors involved. There are many heuristics that are used in practice, but we know of no bounds on their performance. We consider the question of whether there exists a simple and practical approximation algorithm for kmeans clustering. We present a local improvement heuristic based on swapping centers in and out. We prove that this yields a (9 + ε)approximation algorithm. We present an example showing that any approach based on performing a fixed number of swaps achieves an approximation factor of at least (9 − ε) in all sufficiently high dimensions. Thus, our approximation factor is almost tight for algorithms based on performing a fixed number of swaps. To establish the practical value of the heuristic, we present an empirical study that shows that, when combined with
Approximate Nearest Neighbor Queries Revisited
, 1998
"... This paper proposes new methods to answer approximate nearest neighbor queries on a set of n points in ddimensional Euclidean space. For any fixed constant d, a data structure with O(" (1\Gammad)=2 n log n) preprocessing time and O(" (1\Gammad)=2 log n) query time achieves approximat ..."
Abstract

Cited by 62 (4 self)
 Add to MetaCart
This paper proposes new methods to answer approximate nearest neighbor queries on a set of n points in ddimensional Euclidean space. For any fixed constant d, a data structure with O(" (1\Gammad)=2 n log n) preprocessing time and O(" (1\Gammad)=2 log n) query time achieves approximation factor 1 + " for any given 0 ! " ! 1; a variant reduces the "dependence by a factor of " \Gamma1=2 . For any arbitrary d, a data structure with O(d 2 n log n) preprocessing time and O(d 2 log n) query time achieves approximation factor O(d 3=2 ). Applications to various proximity problems are discussed. 1 Introduction Let P be a set of n point sites in ddimensional space IR d . In the wellknown post office problem, we want to preprocess P into a data structure so that a site closest to a given query point q (called the nearest neighbor of q) can be found efficiently. Distances are measured under the Euclidean metric. The post office problem has many applications within computational...
A Cost Model for Query Processing in HighDimensional Data Spaces
, 2000
"... During the last decade, multimedia databases have become increasingly important in many application areas such as medicine, CAD, geography or molecular biology. An important research issue in the field of multimedia databases is similarity search in large data sets. Most current approaches addressin ..."
Abstract

Cited by 53 (0 self)
 Add to MetaCart
During the last decade, multimedia databases have become increasingly important in many application areas such as medicine, CAD, geography or molecular biology. An important research issue in the field of multimedia databases is similarity search in large data sets. Most current approaches addressing similarity search use the socalled feature approach which transforms important properties of the stored objects into points of a highdimensional space (feature vectors). Thus, the similarity search is transformed into a neighborhood search in the feature space. For the management of the feature vectors, multidimensional index structures are usually applied. The performance of query processing can be substantially improved by opti...
A Lower Bound on the Complexity of Approximate NearestNeighbor Searching on the Hamming Cube
 In Proc. 31th Annual ACM Symposium on Theory of Computing (STOC’99
, 1999
"... We consider the nearestneighbor problem over the dcube: given a collection of points in {0, 1} d , find the one nearest to a query point (in the L 1 sense). We establish a lower bound of###90 log d/ log log log d) on the worstcase query time. This result holds in the cell probe model with ( ..."
Abstract

Cited by 20 (3 self)
 Add to MetaCart
We consider the nearestneighbor problem over the dcube: given a collection of points in {0, 1} d , find the one nearest to a query point (in the L 1 sense). We establish a lower bound of###90 log d/ log log log d) on the worstcase query time. This result holds in the cell probe model with (any amount of) polynomial storage and wordsize d O(1) . The same lower bound holds for the approximate version of the problem, where the answer may be any point further than the nearest neighbor by a factor as large as 2 #(log d) 1# # , for any fixed # > 0. 1 Introduction For a variety of practical reasons ranging from molecular biology to web searching, nearestneighbor searching has been a focus of attention lately [2][9], [11][21], [26]. In the applications considered, the dimension of the ambient space is usually high, and predictably, classical lines of attack based on space partitioning fail. To overcome the wellknown "curse of dimensionality," it is typical to relax the s...
Modeling HighDimensional Index Structures using Sampling
 In Proc. ACM SIGMOD Int. Conf. on Management of Data
, 2001
"... A large number of index structures for highdimensional data have been proposed previously. In order to tune and compare such index structures, it is vital to have efficient cost prediction techniques for these structures. Previous techniques either assume uniformity of the data or are not applicabl ..."
Abstract

Cited by 15 (4 self)
 Add to MetaCart
(Show Context)
A large number of index structures for highdimensional data have been proposed previously. In order to tune and compare such index structures, it is vital to have efficient cost prediction techniques for these structures. Previous techniques either assume uniformity of the data or are not applicable to highdimensional data. We propose the use of sampling to predict the number of accessed index pages during a query execution. Sampling is independent of the dimensionality and preserves clusters which is important for representing skewed data. We present a general model for estimating the index page layout using sampling and show how to compensate for errors. We then give an implementation of our model under restricted memory assumptions and show that it performs well even under these constraints. Errors are minimal and the overall prediction time is up to two orders of magnitude below the time for building and probing the full index without sampling. 1.