Deep Hashing with Hash-Consistent Large Margin Proxy Embeddings

TL;DR

This paper introduces hash-consistent large margin proxy embeddings (HCLM) that improve binary hashing performance by eliminating rotational ambiguity and encouraging saturation, with a semantic extension (sHCLM) for transfer scenarios, outperforming state-of-the-art methods.

Contribution

The paper proposes a novel proxy embedding design that reduces binarization error and enhances discriminability, along with a semantic extension for transfer learning, advancing image hashing techniques.

Findings

sHCLM significantly outperforms existing hashing methods on multiple datasets.

HCLM proxies ensure low binarization error and high discriminability.

The approach is effective both within and beyond training classes.

Abstract

Image hash codes are produced by binarizing the embeddings of convolutional neural networks (CNN) trained for either classification or retrieval. While proxy embeddings achieve good performance on both tasks, they are non-trivial to binarize, due to a rotational ambiguity that encourages non-binary embeddings. The use of a fixed set of proxies (weights of the CNN classification layer) is proposed to eliminate this ambiguity, and a procedure to design proxy sets that are nearly optimal for both classification and hashing is introduced. The resulting hash-consistent large margin (HCLM) proxies are shown to encourage saturation of hashing units, thus guaranteeing a small binarization error, while producing highly discriminative hash-codes. A semantic extension (sHCLM), aimed to improve hashing performance in a transfer scenario, is also proposed. Extensive experiments show that sHCLM embeddings achieve significant improvements over state-of-the-art hashing procedures on several small and large datasets, both within and beyond the set of training classes.