A Modification-Free Steganography Algorithm Based on Image Classification and CNN | IGI Global Scientific Publishing
Welcome to the InfoSci Platform
Reference Hub3
A Modification-Free Steganography Algorithm Based on Image Classification and CNN

A Modification-Free Steganography Algorithm Based on Image Classification and CNN

Jian Bin Wu, Yang Zhang, Chu Wei Luo, Lin Feng Yuan, Xiao Kang Shen
Copyright: © 2021 |Volume: 13 |Issue: 3 |Pages: 12
ISSN: 1941-6210|EISSN: 1941-6229|EISBN13: 9781799860365|DOI: 10.4018/IJDCF.20210501.oa4
Cite Article Cite Article

MLA

Wu, Jian Bin, et al. "A Modification-Free Steganography Algorithm Based on Image Classification and CNN." IJDCF vol.13, no.3 2021: pp.47-58. https://doi.org/10.4018/IJDCF.20210501.oa4

APA

Wu, J. B., Zhang, Y., Luo, C. W., Yuan, L. F., & Shen, X. K. (2021). A Modification-Free Steganography Algorithm Based on Image Classification and CNN. International Journal of Digital Crime and Forensics (IJDCF), 13(3), 47-58. https://doi.org/10.4018/IJDCF.20210501.oa4

Chicago

Wu, Jian Bin, et al. "A Modification-Free Steganography Algorithm Based on Image Classification and CNN," International Journal of Digital Crime and Forensics (IJDCF) 13, no.3: 47-58. https://doi.org/10.4018/IJDCF.20210501.oa4

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

In order to improve the data-embedding capacity of modification-free steganography algorithm, scholars have done a lot of research work to meet practical demands. By researching the user's behavioral habits of several social platforms, a semi-structured modification-free steganography algorithm is introduced in the paper. By constructing the mapping relationship between small icons and binary numbers, the idea of image stitching is utilized, and small icons are stitched together according to the behavioral habits of people's social platforms to implement the graphical representation of secret messages. The convolutional neural network (CNN) has been used to train the small icon recognition and classification data set in the algorithm. In order to improve the robustness of the algorithm, the icons processed by various attack methods are introduced as interference samples in the training set. The experimental results show that the algorithm has good anti-attack ability, and the hiding capacity can be improved, which can be used in the covert communication.