Product Quantization of high dimensional vector

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 sub spaces 

that are then quantized separately[1].

In mathematics, a Cartesian product (or product set) is the direct product of two sets. 

Specifically, the Cartesian product of two sets X (for example the points on an x-axis) and Y (for example the points on a y-axis), 

denoted X × Y, is the set of all possible ordered pairs whose first component is a member of X and whose second component

 is a member of Y (e.g., the whole of the x–y plane)[2]

[1] http://research.microsoft.com/pubs/187499/cvpr13opq.pdf

[2] http://www.princeton.edu/~achaney/tmve/wiki100k/docs/Cartesian_product.html