Unsupervised Image Steganalysis Method Using Self-Learning Ensemble Discriminant Clustering

Bing CAO; Guorui FENG; Zhaoxia YIN; Lingyan FAN

doi:10.1587/transinf.2017EDL8011

Unsupervised Image Steganalysis Method Using Self-Learning Ensemble Discriminant Clustering

Bing CAO, Guorui FENG, Zhaoxia YIN, Lingyan FAN

Full Text Views

0

Share
Cite this

Summary :

Image steganography is a technique of embedding secret message into a digital image to securely send the information. In contrast, steganalysis focuses on detecting the presence of secret messages hidden by steganography. The modern approach in steganalysis is based on supervised learning where the training set must include the steganographic and natural image features. But if a new method of steganography is proposed, and the detector still trained on existing methods will generally lead to the serious detection accuracy drop due to the mismatch between training and detecting steganographic method. In this paper, we just attempt to process unsupervised learning problem and propose a detection model called self-learning ensemble discriminant clustering (SEDC), which aims at taking full advantage of the statistical property of the natural and testing images to estimate the optimal projection vector. This method can adaptively select the most discriminative subspace and then use K-means clustering to generate the ultimate class labels. Experimental results on J-UNIWARD and nsF5 steganographic methods with three feature extraction methods such as CC-JRM, DCTR, GFR show that the proposed scheme can effectively classification better than blind speculation.

Publication: IEICE TRANSACTIONS on Information Vol.E100-D No.5 pp.1144-1147

Publication Date: 2017/05/01

Publicized: 2017/02/18

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2017EDL8011

Type of Manuscript: LETTER

Category: Image Recognition, Computer Vision

Authors

Bing CAO
  Shanghai University
Guorui FENG
  Shanghai University
Zhaoxia YIN
  Anhui University
Lingyan FAN
  Hangzhou Dianzi University

Keyword

image steganalysis, statistical property, clustering, unsupervised learning

Cite this

Copy

Bing CAO, Guorui FENG, Zhaoxia YIN, Lingyan FAN, "Unsupervised Image Steganalysis Method Using Self-Learning Ensemble Discriminant Clustering" in IEICE TRANSACTIONS on Information, vol. E100-D, no. 5, pp. 1144-1147, May 2017, doi: 10.1587/transinf.2017EDL8011.
Abstract: Image steganography is a technique of embedding secret message into a digital image to securely send the information. In contrast, steganalysis focuses on detecting the presence of secret messages hidden by steganography. The modern approach in steganalysis is based on supervised learning where the training set must include the steganographic and natural image features. But if a new method of steganography is proposed, and the detector still trained on existing methods will generally lead to the serious detection accuracy drop due to the mismatch between training and detecting steganographic method. In this paper, we just attempt to process unsupervised learning problem and propose a detection model called self-learning ensemble discriminant clustering (SEDC), which aims at taking full advantage of the statistical property of the natural and testing images to estimate the optimal projection vector. This method can adaptively select the most discriminative subspace and then use K-means clustering to generate the ultimate class labels. Experimental results on J-UNIWARD and nsF5 steganographic methods with three feature extraction methods such as CC-JRM, DCTR, GFR show that the proposed scheme can effectively classification better than blind speculation.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2017EDL8011/_p

Copy

@ARTICLE{e100-d_5_1144,
author={Bing CAO, Guorui FENG, Zhaoxia YIN, Lingyan FAN, },
journal={IEICE TRANSACTIONS on Information},
title={Unsupervised Image Steganalysis Method Using Self-Learning Ensemble Discriminant Clustering},
year={2017},
volume={E100-D},
number={5},
pages={1144-1147},
abstract={Image steganography is a technique of embedding secret message into a digital image to securely send the information. In contrast, steganalysis focuses on detecting the presence of secret messages hidden by steganography. The modern approach in steganalysis is based on supervised learning where the training set must include the steganographic and natural image features. But if a new method of steganography is proposed, and the detector still trained on existing methods will generally lead to the serious detection accuracy drop due to the mismatch between training and detecting steganographic method. In this paper, we just attempt to process unsupervised learning problem and propose a detection model called self-learning ensemble discriminant clustering (SEDC), which aims at taking full advantage of the statistical property of the natural and testing images to estimate the optimal projection vector. This method can adaptively select the most discriminative subspace and then use K-means clustering to generate the ultimate class labels. Experimental results on J-UNIWARD and nsF5 steganographic methods with three feature extraction methods such as CC-JRM, DCTR, GFR show that the proposed scheme can effectively classification better than blind speculation.},
keywords={},
doi={10.1587/transinf.2017EDL8011},
ISSN={1745-1361},
month={May},}

Copy

TY - JOUR
TI - Unsupervised Image Steganalysis Method Using Self-Learning Ensemble Discriminant Clustering
T2 - IEICE TRANSACTIONS on Information
SP - 1144
EP - 1147
AU - Bing CAO
AU - Guorui FENG
AU - Zhaoxia YIN
AU - Lingyan FAN
PY - 2017
DO - 10.1587/transinf.2017EDL8011
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2017
AB - Image steganography is a technique of embedding secret message into a digital image to securely send the information. In contrast, steganalysis focuses on detecting the presence of secret messages hidden by steganography. The modern approach in steganalysis is based on supervised learning where the training set must include the steganographic and natural image features. But if a new method of steganography is proposed, and the detector still trained on existing methods will generally lead to the serious detection accuracy drop due to the mismatch between training and detecting steganographic method. In this paper, we just attempt to process unsupervised learning problem and propose a detection model called self-learning ensemble discriminant clustering (SEDC), which aims at taking full advantage of the statistical property of the natural and testing images to estimate the optimal projection vector. This method can adaptively select the most discriminative subspace and then use K-means clustering to generate the ultimate class labels. Experimental results on J-UNIWARD and nsF5 steganographic methods with three feature extraction methods such as CC-JRM, DCTR, GFR show that the proposed scheme can effectively classification better than blind speculation.
ER -