Implicit Sparse Code Hashing

TL;DR

This paper introduces a novel unsupervised binary coding method that preserves inner product relationships of sparse codes for efficient approximate nearest-neighbor search in high-dimensional image data.

Contribution

It proposes a sparse code hashing approach that avoids explicit sparse coding and addresses eigenproblem complexity through column sampling and rotation matrix assumptions.

Findings

Outperforms state-of-the-art binary coding methods on image datasets

Effectively preserves image relationships via inner product in sparse codes

Reduces computational complexity with column sampling and rotation assumptions

Abstract

We address the problem of converting large-scale high-dimensional image data into binary codes so that approximate nearest-neighbor search over them can be efficiently performed. Different from most of the existing unsupervised approaches for yielding binary codes, our method is based on a dimensionality-reduction criterion that its resulting mapping is designed to preserve the image relationships entailed by the inner products of sparse codes, rather than those implied by the Euclidean distances in the ambient space. While the proposed formulation does not require computing any sparse codes, the underlying computation model still inevitably involves solving an unmanageable eigenproblem when extremely high-dimensional descriptors are used. To overcome the difficulty, we consider the column-sampling technique and presume a special form of rotation matrix to facilitate subproblem decomposition. We test our method on several challenging image datasets and demonstrate its effectiveness by comparing with state-of-the-art binary coding techniques.