Uniform Variance Product Quantization

Qin Zhen Guo; Zhi Zeng; Shu Wu Zhang

doi:10.4028/www.scientific.net/AMM.651-653.2224

Paper Titles

Sufficient Conditions of H-Matrix
p.2207

The Equivalence Relationship of Matrix and the Corresponding Equivalence Classes
p.2211

The Identity between the Real Number and Order one Matrix
p.2216

The Image Fusion Algorithm Based on Rough Neural Network
p.2220

Uniform Variance Product Quantization
p.2224

Using Sample and Multilayer Compressed Counting Bloom Filter Algorithm to Realize Elephant Flows Identification
p.2228

Simulation and Analysis on Fast Modeling for Large-Scale 3D Dynamic Images
p.2233

SIMP-Based Evolutionary Structural Optimization Method for Topology Optimization
p.2237

The New Approach to the Invariant Feature Extraction Using Ridgelet Transform
p.2241

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vols. 651-653Uniform Variance Product Quantization

Uniform Variance Product Quantization

Abstract:

Product quantization (PQ) is an efficient and effective vector quantization approach to fast approximate nearest neighbor (ANN) search especially for high-dimensional data. The basic idea of PQ is to decompose the original data space into the Cartesian product of some low-dimensional subspaces and then every subspace is quantized separately with the same number of codewords. However, the performance of PQ depends largely on the distribution of the original data. If the distributions of every subspace have larger difference, PQ will achieve bad results as shown in our experiments. In this paper, we propose a uniform variance product quantization (UVPQ) scheme to project the data by a uniform variance projection before decompose it, which can minimize the subspace distribution difference of the whole space. UVPQ can guarantee good results however the data rotate. Extensive experiments have verified the superiority of UVPQ over PQ for ANN search.