Bi-level Locality Sensitive Hashing for k-Nearest Neighbor Computation

We present a new Bi-level LSH algorithm to perform approximate k-nearest neighbor search in high dimensional spaces. Our formulation is based on a two-level scheme. In the first level, we use a RP-tree that divides the dataset into subgroups with bounded aspect ratios and is used to distinguish well-separated clusters. During the second level, we construct one LSH hash table for each sub-group, which is enhanced by a hierarchical structure based on space-filling curves and lattice techniques. Given a query, we first determine the sub-group that it belongs to and then perform a k-nearest neighbor search within the suitable buckets in the LSH hash table corresponding to the sub-group. In practice, our algorithm is able to improve the quality and reduce the runtime of approximate k-nearest neighbor computations. We demonstrate the performance of our method on two large, high-dimensional and widely used image datasets and show that when given the same runtime budget, our bi-level metho...

• Type: Journal
• Year: 2012
• Where: ICDE
• Authors: Jia Pan, Dinesh Manocha
• Keywords: Aspect Ratios, Database, ICDE 2012, Image Datasets, Neighbor Search