SteganoCNN: Image Steganography with Generalization Ability Based on Convolutional Neural Network.

Nao Liu,Wenxin Wang,Mengxiao Gou,Xintao Duan,Chuan Qin

doi:10.3390/e22101140

Abstract

Image-to-image steganography is hiding one image in another image. However, hiding two secret images into one carrier image is a challenge today. The application of image steganography based on deep learning in real-life is relatively rare. In this paper, a new Steganography Convolution Neural Network (SteganoCNN) model is proposed, which solves the problem of two images embedded in a carrier image and can effectively reconstruct two secret images. SteganoCNN has two modules, an encoding network, and a decoding network, whereas the decoding network includes two extraction networks. First, the entire network is trained end-to-end, the encoding network automatically embeds the secret image into the carrier image, and the decoding network is used to reconstruct two different secret images. The experimental results show that the proposed steganography scheme has a maximum image payload capacity of 47.92 bits per pixel, and at the same time, it can effectively avoid the detection of steganalysis tools while keeping the stego-image undistorted. Meanwhile, StegaoCNN has good generalization capabilities and can realize the steganography of different data types, such as remote sensing images and aerial images.

Highlights

Privacy protection of communication between two parties has been a hot topic on the internet for a long time, and the privacy protection of communication involved [1] can be combined with information hiding, thereby achieving secure communication [2]
We propose a steganography model (SteganoCNN) based on a deep convolutional neural network that can effectively increase the steganography capacity under the premise of ensuring steganography security
The final result of the experiment is analyzed through visual and quantitative evaluation, and the safety of steganography is checked by steganalysis tools

Summary

Introduction

Privacy protection of communication between two parties has been a hot topic on the internet for a long time, and the privacy protection of communication involved [1] can be combined with information hiding, thereby achieving secure communication [2]. The secret information is embedded in the disturbing image, and the disturbing image is sent to the GAN to continue training and output as stego. This is sent to the recipient, who extracts the secret information through the Cardan grille mask agreed upon by both parties.

Related Work

SteganoCNN Architecture

Encoder That Hides Secret Images

Ordinary CNN

FCDenseNet

Decoder for Reconstructing the Secret Image

Experimental Results and Analysis

Subjective Visual Assessment

Peak Signal-to-Noise Ratio and Structural Similarity

Relative Capacity and Payload Capacity

Generalization Ability

Steganalysis