Zero-Shot Hashing via Transferring Supervised Knowledge

TL;DR

This paper introduces zero-shot hashing (ZSH), a method that transfers supervised knowledge to generate hash codes for unseen categories, improving large-scale multimedia retrieval without extensive manual labeling.

Contribution

The paper proposes a novel zero-shot hashing framework that transfers semantic knowledge to unseen classes and aligns semantic and visual spaces to enhance hashing performance.

Findings

ZSH outperforms state-of-the-art hashing methods in zero-shot image retrieval.

The semantic space rotation improves alignment with visual features.

Efficient alternating algorithm enables practical application of ZSH.

Abstract

Hashing has shown its efficiency and effectiveness in facilitating large-scale multimedia applications. Supervised knowledge e.g. semantic labels or pair-wise relationship) associated to data is capable of significantly improving the quality of hash codes and hash functions. However, confronted with the rapid growth of newly-emerging concepts and multimedia data on the Web, existing supervised hashing approaches may easily suffer from the scarcity and validity of supervised information due to the expensive cost of manual labelling. In this paper, we propose a novel hashing scheme, termed \emph{zero-shot hashing} (ZSH), which compresses images of "unseen" categories to binary codes with hash functions learned from limited training data of "seen" categories. Specifically, we project independent data labels i.e. 0/1-form label vectors) into semantic embedding space, where semantic relationships among all the labels can be precisely characterized and thus seen supervised knowledge can be transferred to unseen classes. Moreover, in order to cope with the semantic shift problem, we rotate the embedded space to more suitably align the embedded semantics with the low-level visual feature space, thereby alleviating the influence of semantic gap. In the meantime, to exert positive effects on learning high-quality hash functions, we further propose to preserve local structural property and discrete nature in binary codes. Besides, we develop an efficient alternating algorithm to solve the ZSH model. Extensive experiments conducted on various real-life datasets show the superior zero-shot image retrieval performance of ZSH as compared to several state-of-the-art hashing methods.