Supervised Quantization for Similarity Search

TL;DR

This paper introduces supervised quantization, a method that learns discriminative low-dimensional representations and quantizes data for efficient, semantically-aware image similarity search, outperforming existing methods.

Contribution

It proposes a novel supervised quantization technique that combines feature selection, low-dimensional transformation, and quantization optimized for semantic separability.

Findings

Outperforms state-of-the-art supervised hashing methods.

Achieves more accurate semantic clustering in low-dimensional space.

Demonstrates superior search accuracy on standard datasets.

Abstract

In this paper, we address the problem of searching for semantically similar images from a large database. We present a compact coding approach, supervised quantization. Our approach simultaneously learns feature selection that linearly transforms the database points into a low-dimensional discriminative subspace, and quantizes the data points in the transformed space. The optimization criterion is that the quantized points not only approximate the transformed points accurately, but also are semantically separable: the points belonging to a class lie in a cluster that is not overlapped with other clusters corresponding to other classes, which is formulated as a classification problem. The experiments on several standard datasets show the superiority of our approach over the state-of-the art supervised hashing and unsupervised quantization algorithms.