Detection of audio covert channels using statistical footprints of hidden messages

Hamza Özer, Bülent Sankur, Nasir Memon, Ismail Avcibaş

Research output: Contribution to journalArticle

Abstract

We address the problem of detecting the presence of hidden messages in audio. The detector is based on the characteristics of the denoised residuals of the audio file, which may consist of a mixture of speech and music data. A set of generalized moments of the audio signal is measured in terms of objective and perceptual quality measures. The detector discriminates between cover and stego files using a selected subset of features and an SVM classifier. The proposed scheme achieves on the average 88% discrimination performance on individual steganographic algorithms and 98.5% on individual watermarking algorithms. Between 75 and 90% discrimination performance is achieved in universal tests. Correct detection performance for individual embedding algorithms is roughly 90% when the detector can encounter any one in an ensemble of different embedding algorithms.

Original languageEnglish (US)
Pages (from-to)389-401
Number of pages13
JournalDigital Signal Processing: A Review Journal
Volume16
Issue number4
DOIs
StatePublished - Jul 2006

Fingerprint

Detectors
Watermarking
Classifiers

Keywords

  • Feature selection
  • Steganalysis
  • Support vector machine
  • Watermarking

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Detection of audio covert channels using statistical footprints of hidden messages. / Özer, Hamza; Sankur, Bülent; Memon, Nasir; Avcibaş, Ismail.

In: Digital Signal Processing: A Review Journal, Vol. 16, No. 4, 07.2006, p. 389-401.

Research output: Contribution to journalArticle

Özer, Hamza ; Sankur, Bülent ; Memon, Nasir ; Avcibaş, Ismail. / Detection of audio covert channels using statistical footprints of hidden messages. In: Digital Signal Processing: A Review Journal. 2006 ; Vol. 16, No. 4. pp. 389-401.
@article{145df51c55564cb781eb4e8ebf571add,
title = "Detection of audio covert channels using statistical footprints of hidden messages",
abstract = "We address the problem of detecting the presence of hidden messages in audio. The detector is based on the characteristics of the denoised residuals of the audio file, which may consist of a mixture of speech and music data. A set of generalized moments of the audio signal is measured in terms of objective and perceptual quality measures. The detector discriminates between cover and stego files using a selected subset of features and an SVM classifier. The proposed scheme achieves on the average 88{\%} discrimination performance on individual steganographic algorithms and 98.5{\%} on individual watermarking algorithms. Between 75 and 90{\%} discrimination performance is achieved in universal tests. Correct detection performance for individual embedding algorithms is roughly 90{\%} when the detector can encounter any one in an ensemble of different embedding algorithms.",
keywords = "Feature selection, Steganalysis, Support vector machine, Watermarking",
author = "Hamza {\"O}zer and B{\"u}lent Sankur and Nasir Memon and Ismail Avcibaş",
year = "2006",
month = "7",
doi = "10.1016/j.dsp.2005.12.001",
language = "English (US)",
volume = "16",
pages = "389--401",
journal = "Digital Signal Processing: A Review Journal",
issn = "1051-2004",
publisher = "Elsevier Inc.",
number = "4",

}

TY - JOUR

T1 - Detection of audio covert channels using statistical footprints of hidden messages

AU - Özer, Hamza

AU - Sankur, Bülent

AU - Memon, Nasir

AU - Avcibaş, Ismail

PY - 2006/7

Y1 - 2006/7

N2 - We address the problem of detecting the presence of hidden messages in audio. The detector is based on the characteristics of the denoised residuals of the audio file, which may consist of a mixture of speech and music data. A set of generalized moments of the audio signal is measured in terms of objective and perceptual quality measures. The detector discriminates between cover and stego files using a selected subset of features and an SVM classifier. The proposed scheme achieves on the average 88% discrimination performance on individual steganographic algorithms and 98.5% on individual watermarking algorithms. Between 75 and 90% discrimination performance is achieved in universal tests. Correct detection performance for individual embedding algorithms is roughly 90% when the detector can encounter any one in an ensemble of different embedding algorithms.

AB - We address the problem of detecting the presence of hidden messages in audio. The detector is based on the characteristics of the denoised residuals of the audio file, which may consist of a mixture of speech and music data. A set of generalized moments of the audio signal is measured in terms of objective and perceptual quality measures. The detector discriminates between cover and stego files using a selected subset of features and an SVM classifier. The proposed scheme achieves on the average 88% discrimination performance on individual steganographic algorithms and 98.5% on individual watermarking algorithms. Between 75 and 90% discrimination performance is achieved in universal tests. Correct detection performance for individual embedding algorithms is roughly 90% when the detector can encounter any one in an ensemble of different embedding algorithms.

KW - Feature selection

KW - Steganalysis

KW - Support vector machine

KW - Watermarking

UR - http://www.scopus.com/inward/record.url?scp=33745384504&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745384504&partnerID=8YFLogxK

U2 - 10.1016/j.dsp.2005.12.001

DO - 10.1016/j.dsp.2005.12.001

M3 - Article

AN - SCOPUS:33745384504

VL - 16

SP - 389

EP - 401

JO - Digital Signal Processing: A Review Journal

JF - Digital Signal Processing: A Review Journal

SN - 1051-2004

IS - 4

ER -