Zero-shot Hashing with Orthogonal Projection for Image Retrieval

Zhang, Haofeng, Long, Yang and Shao, Ling (2019) Zero-shot Hashing with Orthogonal Projection for Image Retrieval. Pattern Recognition Letters, 117. pp. 201-209. ISSN 0167-8655

[img] PDF (Accepted manuscript) - Submitted Version
Restricted to Repository staff only until 09 April 2019.
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (7MB) | Request a copy

    Abstract

    Hashing has been widely used in large-scale image retrieval. Supervised information such as semantic similarity and class label, and Convolutional Neural Network (CNN) has greatly improved the quality of hash codes and hash functions. However, due to the explosive growth of web data, existing hashing methods can not well perform on emerging images of new classes. In this paper, we propose a novel hashing method based on orthogonal projection of both image and semantic attribute, which constrains the generated binary codes in orthogonal space should be orthogonal with each other when they belong to different classes, otherwise be same. This constraint guarantees that the generated hash codes from different categories have equal Hamming distance, which also makes the space more discriminative within limited code length. To improve the performance, we also extend our method with a deep model. Experiments of both our linear and deep model on three popular datasets show that our method can achieve competitive results, specially, the deep model can outperform all the listed state-of-the-art approaches.

    Item Type: Article
    Uncontrolled Keywords: zero-shot hashing,orthogonal projection,image retrieval
    Faculty \ School: Faculty of Science > School of Computing Sciences
    Related URLs:
    Depositing User: Pure Connector
    Date Deposited: 13 Apr 2018 17:30
    Last Modified: 11 Jan 2019 09:30
    URI: https://ueaeprints.uea.ac.uk/id/eprint/66764
    DOI: 10.1016/j.patrec.2018.04.011

    Actions (login required)

    View Item