A Bidirectional Generative Adversarial Network-Based Perceptual Hash Algorithm for Image Content Forensics

Document Type

Article

Publication Date

12-1-2023

Abstract

The traditional perceptual hash algorithm creates image perceptual hash code by extracting image features with a pre-designed scheme. As it is hard to make full use of image inherent semantic characters, the performance of perceptual hash code on image content authentication and copyright protection is constrained. In this paper, an unsupervised perceptual hash algorithm for image forensics based on Bidirectional Generative Adversarial Network （BiGAN） is proposed. The main contributions of the paper are as follows:Firstly, depending on the bidirectional iterative adversary among the coding network, the generative network, and the discriminative network, the powerful learning ability of BiGAN on image inherent feature extraction is fully developed;so that the perceptual hash code that has strong image semantic feature representation capability can be created. As a result, both the identification robustness for images with identical content and the discrimination sensitivity for images with different contents are achieved. Hence, the capability of image forensics is improved. Secondly, a BiGAN optimization framework is constructed by adding a skip-connection structure between the coding and the generative network. By concatenating the shallow and deep layers′ features of the sampled image, different dimensional features are organically integrated to improve the learning efficiency and the convergence speed of the proposed scheme. Thereby, the semantic information representation ability of the perceptual hash code is enhanced, and the identification robustness for identical content images is heightened. Thirdly, a Mean Square Error （MSE） loss-based performance optimization strategy for BiGAN is investigated. By computing the difference between the output of the coding network and the generative network, not only the visual quality of the generated image but also the representation capability of the generated perceptual hash code is effectively improved. Consequently, the discrimination sensitivity for different content images is intensified. In the end, by virtue of multiple network iterations and adversarial training, a high-performance perceptual hash code for image forensics is obtained. Furthermore, a large image database CeleA Mask-HQ is employed for the first time to evaluate the performance of the perceptual hash algorithm in this study. The capability of the BiGAN-based perceptual hash algorithm for the identification of images with identical content and for the discrimination of images with different contents is discussed in detail. Meanwhile, both the influence of the skip-connection network structure and that of the mean square error （MSE） loss on the performance improvement of the BiGAN-based perceptual hash algorithm are explored at length. In addition, four excellent image perceptual hash algorithms are involved in the experiment to verify the performance of the proposed scheme in comparisons. Extensive experimental results indicate that the BiGAN-based perceptual hash algorithm gains higher image forensics ability than other state-of-the-art schemes.

Identifier

85179755999 (Scopus)

Publication Title

Jisuanji Xuebao Chinese Journal of Computers

External Full Text Location

https://doi.org/10.11897/SP.J.1016.2023.02551

ISSN

02544164

First Page

2551

Last Page

2572

Issue

12

Volume

46

This document is currently not available here.

Share

COinS