Reversible data hiding in shared JPEG images for distributed framework

Abstract

Reversible Data Hiding in Encrypted Images (RDHEI) has emerged as an important technique for securely transmitting sensitive information over open networks. However, existing research suffers from two main limitations. Most schemesrely on a single data hider, which introduces the risk of a single point of failure, and they are primarily designed for uncompressed images, making them unsuitable for JPEG images. To address the secure embedding needs of encrypted images in distributed environments, we propose a reversible data hiding in shared JPEG images for distributed framework. A secure multi-party embedding framework is constructed wherein both the original JPEG image and the secret data are partitioned into multiple shares and distributed to different embedders, enabling fully distributed data embedding. Within this framework, the DC coefficients are protected using iterative encryption, while AC coefficients and secret data are processed through a polynomial secret sharing mechanism based on Hamming codes. Data embedding is implemented by means of XOR operations, all of which are executed directly on the JPEG bitstream. Even if some image shares are lost during transmission or storage, the secret data can be fully extracted and the original JPEG image can be losslessly recovered using the remaining marked JPEG shares. Experimental results show that the proposed method is highly compatible with the JPEG format, effectively preserves the file size of each JPEG share, and achieves higher embedding capacity compared to existing methods.