The pyramid-technique: towards breaking the curse of dimensionality
SIGMOD '98 Proceedings of the 1998 ACM SIGMOD international conference on Management of data
R-trees: a dynamic index structure for spatial searching
SIGMOD '84 Proceedings of the 1984 ACM SIGMOD international conference on Management of data
VLDB '98 Proceedings of the 24rd International Conference on Very Large Data Bases
Efficient similarity search and classification via rank aggregation
Proceedings of the 2003 ACM SIGMOD international conference on Management of data
Contorting high dimensional data for efficient main memory KNN processing
Proceedings of the 2003 ACM SIGMOD international conference on Management of data
Towards effective indexing for very large video sequence database
Proceedings of the 2005 ACM SIGMOD international conference on Management of data
iDistance: An adaptive B+-tree based indexing method for nearest neighbor search
ACM Transactions on Database Systems (TODS)
Near-Optimal Hashing Algorithms for Approximate Nearest Neighbor in High Dimensions
FOCS '06 Proceedings of the 47th Annual IEEE Symposium on Foundations of Computer Science
Access Structures for Angular Similarity Queries
IEEE Transactions on Knowledge and Data Engineering
Multidimensional Binary Search Trees in Database Applications
IEEE Transactions on Software Engineering
Near-optimal hashing algorithms for approximate nearest neighbor in high dimensions
Communications of the ACM - 50th anniversary issue: 1958 - 2008
Quality and efficiency in high dimensional nearest neighbor search
Proceedings of the 2009 ACM SIGMOD International Conference on Management of data
An efficient indexing method for nearest neighbor searches inhigh-dirnensional image databases
IEEE Transactions on Multimedia
Enhancing minimum spanning tree-based clustering by removing density-based outliers
Digital Signal Processing
Hi-index | 0.00 |
K-nearest neighbor (KNN) search in high dimensional space is essential for database applications, especially multimedia database applications, because images and audio clips are always modeled as high dimensional vectors. However, performance of existing indexing methods degrades dramatically as the dimensionality increases. In this paper, we propose a novel polar coordinate based indexing method, called iPoc, for efficient KNN search in high dimensional space. First, data space is initially partitioned into hypersphere regions, and then each hypersphere is further refined into hypersectors via hyperspherical surface clustering. After that, a series of local polar coordinate systems can be derived from hypersectors, taking advantage of the geometric characters of hypersectors. During search processing, iPoc can effectively prune query-unrelated data points by estimating the lower and upper bounds in both radial coordinate and angle coordinate. Furthermore, we design a key mapping scheme to merge keys measured by independent local polar coordinates into the global polar coordinates. Finally, the global polar coordinates are indexed by a traditional 2-dimensional spatial index, e.g., R-tree. Extensive experiments on real audio datasets and synthetic datasets confirm the effectiveness and efficiency of our proposal and prove that iPoc is more efficient than the existing high dimensional KNN search methods.