Abstract
The technique of reversible data hiding in encrypted images (RDH-EI) has experienced significant interest as it allows for precise extraction of embedded data without compromising the confidentiality of the original image. This research introduces a novel RDH-EI technique designed to accommodate multiple data hiders. To tackle this challenge, we propose a sophisticated RDH-EI method that integrates secret sharing Founded on the Learning With Errors (LWE) problem alongside adaptive coding strategies. On the content owner's side, the original image is first distributed to multiple data hiders using a method Founded on the Learning With Errors (LWE) problem. Then, block permutation along with stream cipher encryption are performed to completely preserve the spatial correlation between image blocks. The proposed method benefits from the robust security provided by LWE. Initially, we examine the spread of the most significant bit planes to detect segments that are suitable for data embedding. Next, the data hider produces extra data and embeds errors within the encrypted image to guarantee precise image reconstruction. To increase data storage capacity, the most significant bits (MSB) The blocks suitable for embedding are adaptively compressed based on their frequency of occurrence. The extra data may be inserted into the (MSB) of the encrypted image, where it is combined with inverse Huffman codewords and supplementary auxiliary information. At the receiving side, the initial image can still be completely restored Losslessly, even if some shares happen to be damaged or missing, provided that a sufficient number of valid shares are available. Experimental findings demonstrate that the RDH-EI approach exceeds the performance of various cutting-edge methods, including those employing secret sharing (SS), in terms of embedding capacity.