BSS: Boosted steganography scheme with cover image preprocessing

作者:

Highlights:

摘要

The existing powerful steganalyzers can find out the presence of secret information in images with high accuracy. Increasing the embedding capacity of cover images reduces the detection risk of stego images. In this respect, we introduce boosted steganography scheme (BSS) that has a preprocessing stage before applying steganography methods. The goal of BSS is increasing the undetectability of stego images. Due to the dependence of embedding capacity of images to their content, we proposed an ensemble steganalyzer to estimate the embedding capacity of each cover image. Since the content of cover images has less significance in steganography, therefore to have more security, the steganographer can select a cover image from a database to achieve higher security and satisfactory embedding capacity. We present several experiments that show the effectiveness of boosted steganography scheme in improving the security of stego images. The experimental results demonstrate that considering a preprocessing stage can significantly improve the steganography security.

论文关键词:Steganalysis,Steganography capacity,Support Vector Machine (SVM),Ensemble methods

论文评审过程:Available online 11 May 2010.

论文官网地址:https://doi.org/10.1016/j.eswa.2010.04.071