Optimized Product Quantization for Approximate Nearest Neighbor Search

Product quantization is an effective vector quantization approach to compactly encode high-dimensional vectors for fast approximate nearest neighbor (ANN) search. The essence of product quantization is to decompose the original high-dimensional space into the Cartesian product of a finite number of...

Full description

Saved in:
Bibliographic Details
Published in2013 IEEE Conference on Computer Vision and Pattern Recognition pp. 2946 - 2953
Main Authors Tiezheng Ge, Kaiming He, Qifa Ke, Jian Sun
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2013
Subjects
Artificial neural networks
Eigenvalues and eigenfunctions
Encoding
Linear programming
nearest neighbor search
Optimization
product quantization
Quantization (signal)
Vectors
Online AccessGet full text

Cover

More Information
Summary:Product quantization is an effective vector quantization approach to compactly encode high-dimensional vectors for fast approximate nearest neighbor (ANN) search. The essence of product quantization is to decompose the original high-dimensional space into the Cartesian product of a finite number of low-dimensional subspaces that are then quantized separately. Optimal space decomposition is important for the performance of ANN search, but still remains unaddressed. In this paper, we optimize product quantization by minimizing quantization distortions w.r.t. the space decomposition and the quantization codebooks. We present two novel methods for optimization: a non-parametric method that alternatively solves two smaller sub-problems, and a parametric method that is guaranteed to achieve the optimal solution if the input data follows some Gaussian distribution. We show by experiments that our optimized approach substantially improves the accuracy of product quantization for ANN search.
ISSN:1063-6919
1063-6919
DOI:10.1109/CVPR.2013.379