Towards Optimal Discrete Online Hashing with Balanced Similarity

TL;DR

This paper introduces BSODH, a novel online hashing method that directly optimizes discrete codes by balancing similarity, effectively preserving data relationships and outperforming existing methods on benchmark datasets.

Contribution

The paper proposes a unified framework for online hashing that directly optimizes discrete codes and addresses data imbalance with a balanced similarity measure.

Findings

Outperforms state-of-the-art methods on benchmark datasets

Effectively preserves similarity between streaming and existing data

Enables direct discrete optimization in online hashing

Abstract

When facing large-scale image datasets, online hashing serves as a promising solution for online retrieval and prediction tasks. It encodes the online streaming data into compact binary codes, and simultaneously updates the hash functions to renew codes of the existing dataset. To this end, the existing methods update hash functions solely based on the new data batch, without investigating the correlation between such new data and the existing dataset. In addition, existing works update the hash functions using a relaxation process in its corresponding approximated continuous space. And it remains as an open problem to directly apply discrete optimizations in online hashing. In this paper, we propose a novel supervised online hashing method, termed Balanced Similarity for Online Discrete Hashing (BSODH), to solve the above problems in a unified framework. BSODH employs a well-designed hashing algorithm to preserve the similarity between the streaming data and the existing dataset via an asymmetric graph regularization. We further identify the "data-imbalance" problem brought by the constructed asymmetric graph, which restricts the application of discrete optimization in our problem. Therefore, a novel balanced similarity is further proposed, which uses two equilibrium factors to balance the similar and dissimilar weights and eventually enables the usage of discrete optimizations. Extensive experiments conducted on three widely-used benchmarks demonstrate the advantages of the proposed method over the state-of-the-art methods.