Analysis of Image Steganography Using Cross-Subband Correlation in the Wavelet Domain

[ 31 October 2022 | vol. 15 | no. 2 | pp. 1-10 ]

About Authors:

Majida Meshari
-Tikrit University, Iraq

Abstract:

This paper introduces a novel image steganalysis scheme that focuses on the correlations between interscale and intrascale wavelet subbands. The proposed method begins with the decomposition of the input image into eight distinct pairs of interscale subbands, which are derived through wavelet transformation. From these subbands, we extract joint characteristic function moments as key features that capture the essential information related to hidden content within the image. Furthermore, we enhance the analysis by extracting singular values from the characteristic function moments associated with the co-occurrence values of the intrascale wavelet subbands, all within the same scale. This step allows us to represent structural properties of the image at different levels of detail, thereby improving the robustness of the feature set. The extracted features are then utilized as input for a support vector machine (SVM) classifier, which is trained to differentiate between stego images (images containing hidden information) and cover images (original images without hidden data). Our extensive simulation results demonstrate that the proposed steganalysis approach significantly outperforms earlier methods in terms of detection rates, indicating its effectiveness in identifying steganography in images. This advancement in steganalysis techniques contributes to enhancing security measures and ensuring the integrity of digital multimedia content.

Keywords:

Support Vector Machine, Steganalysis, Interscale and Intrascale Correlations, Characteristic Function; Singular Value

 

About this Article: