A privacy protection scheme for green communication combining digital steganography

In order to ensure the sustainable development of the Internet of Things (IoT), researchers have been exploring and making technological innovations in the fields of smart grid as well as green communication and green computing. In the new communication and computing paradigm, information security and privacy protection of communication is a key research area that is receiving more and more attention. In order to further enhance the covertness of communication methods and privacy protection of communication contents, this paper proposes a user communication scheme based on the digital steganography method, which creates an image to hide the existence of the information itself only according to the secret information to be delivered, and can effectively avoid the image distortion caused by the embedding method. In this scheme, according to the proposed algorithm of interconversion between information and pixels, the secret information is encoded into pixel format and mapped in a blank image, which is considered as a broken image to be repaired. The broken image is repaired by training the network to get the steganographic image. The receiver selects specific steganographic pixels from the received image and further converts them into secret messages. In addition, the proposed communication framework supports different generative network structures that can further improve the steganography resistance. The communication scheme is trained and tested on several datasets and qualitatively and quantitatively evaluated on steganographic images. The final experimental results show that the framework achieves satisfactory performance in terms of both steganography capacity and information recovery rate compared to other embedding-free methods.