A multi-task learning CNN for image steganalysis

Xiangyu Yu, Huabin Tan, Hui Liang, Chang Tsun Li, Guangjun Liao

    Research output: Book chapter/Published conference paperConference paperpeer-review

    10 Citations (Scopus)


    Convolutional neural network (CNN) based image steganalysis are increasingly popular because of their superiority in accuracy. The most straightforward way to employ CNN for image steganalysis is to learn a CNN-based classifier to distinguish whether secret messages have been embedded into an image. However, it is difficult to learn such a classifier because of the weak stego signals and the limited useful information. To address this issue, in this paper, a multi-task learning CNN is proposed. In addition to the typical use of CNN, learning a CNN-based classifier for the whole image, our multi-task CNN is learned with an auxiliary task of the pixel binary classification, estimating whether each pixel in an image has been modified due to steganography. To the best of our knowledge, we are the first to employ CNN to perform the pixel-level classification of such type. Experimental results have justified the effectiveness and efficiency of the proposed multi-task learning CNN.

    Original languageEnglish
    Title of host publication10th IEEE International Workshop on Information Forensics and Security, WIFS 2018
    Place of PublicationUnited States
    PublisherIEEE, Institute of Electrical and Electronics Engineers
    Number of pages7
    ISBN (Electronic)9781538665367
    ISBN (Print)9781538665374 (Print on demand)
    Publication statusPublished - 30 Jan 2019
    EventIEEE International Workshop on Information Forensics and Security: WIFS 2018 - Li Ka Shing Tower, The Hong Kong Polytechnic University, Hung Hom, Hong Kong
    Duration: 10 Dec 201813 Dec 2018
    https://wifs2018.comp.polyu.edu.hk/ (conference website)


    ConferenceIEEE International Workshop on Information Forensics and Security
    Country/TerritoryHong Kong
    CityHung Hom
    OtherIEEE International Workshop on Information Forensics and Security (WIFS) is a unique annual event which is technically co-sponsored by the IEEE Biometrics Council and IEEE Signal Processing Society. The 2018 edition of this workshop will be held in Hong Kong. WIFS is the unique workshop series organised by the IEEE Information Forensics and Security (IFS) Technical Committee of the IEEE Signal Processing Society. It's a major forum that brings researchers from related disciplines to discuss emerging challenges in different areas of information security and forensics, and share the latest results. This workshop serves to provide a high quality forum for advancing research and development efforts in a range of areas defining e-security and forensics.
    The 10th edition of WIFS 2018 will be held in Hong Kong and will feature keynotes, tutorials, special sessions, panel sessions, along with oral and poster sessions.
    Internet address


    Dive into the research topics of 'A multi-task learning CNN for image steganalysis'. Together they form a unique fingerprint.

    Cite this