Year: 2024
Type: conference-paper
Abstract: Nowadays, image encryption and secret sharing are commonly combined to protect the private information of images and to gain disaster-tolerance. However, encrypted images are also susceptible to tampering when transmitted over insecure channels, and secret sharing techniques cannot guarantee tampering localization and lossless recovery when each share of the encrypted image is tampered with. Thus, in this paper, we propose a self-embedding watermarking scheme for ciphertext images, which can realize tampering localization and lossless recovery. In detail, we first design a lossless compression algorithm to compress the original image to obtain a redundant space for embedding authentication bits. Then, we encrypt the compress data and use reference sharing mechanism to generate reference bits for recovering tampered content. In addition, a reversible data hiding scheme based on secret image sharing is applied to divide the encrypted image into multiple ciphertext images, and reference bits are embedded during this process. At the receiver end, integrity authentication and tampering localization are performed on the collected images based on the authentication bits. Then, according to the results of the integrity authentication, the original image can be recovered losslessly through image-based or block-based recovery. Extensive experimental results show that the proposed scheme achieves high performance in terms of recovery quality, security, tampering resistance, etc.
Authors: Haowei Ye, Lin Huang, Lin Lu, Fang Cao, Chuan Qin
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4411729806