Reversible data hiding in encrypted images(RDH-EI)enables the concealment of secret data within ciphertext images while preserving the ability to fully recover both the original image and the hidden message.However,ex...Reversible data hiding in encrypted images(RDH-EI)enables the concealment of secret data within ciphertext images while preserving the ability to fully recover both the original image and the hidden message.However,existing RDH-EI schemes based on vacating room after encryption(VRAE)suffer from limited embedding capacity.To address this issue,we propose a method based on arithmetic coding and dual prediction for encrypted images.First,the original image is encrypted with a chunked modulus and permutation.Then,using the upper-left corner pixel of each subblock as a reference,adaptive MSB prediction and difference prediction are employed to predict the remaining pixels within the subblock.The resulting label map is then compressed via arithmetic coding to vacate the embedding space for the secret message.Finally,the separable operations of the original image restoration and secret message extraction can be performed on the basis of the type of key possessed.The experimental results demonstrate that the proposed algorithm not only successfully extracts the secret information but also recovers the original image without any loss.Furthermore,it effectively enhances the embedding capacity by fully utilizing the correlation between adjacent pixels while ensuring the security of the image.展开更多
基金supported by the Provincial Colleges Quality Project of Anhui Prov-ince(2020xsxxkc047)the National Undergraduate Innovation and Entrepreneurship Training Program(2023103570289).
文摘Reversible data hiding in encrypted images(RDH-EI)enables the concealment of secret data within ciphertext images while preserving the ability to fully recover both the original image and the hidden message.However,existing RDH-EI schemes based on vacating room after encryption(VRAE)suffer from limited embedding capacity.To address this issue,we propose a method based on arithmetic coding and dual prediction for encrypted images.First,the original image is encrypted with a chunked modulus and permutation.Then,using the upper-left corner pixel of each subblock as a reference,adaptive MSB prediction and difference prediction are employed to predict the remaining pixels within the subblock.The resulting label map is then compressed via arithmetic coding to vacate the embedding space for the secret message.Finally,the separable operations of the original image restoration and secret message extraction can be performed on the basis of the type of key possessed.The experimental results demonstrate that the proposed algorithm not only successfully extracts the secret information but also recovers the original image without any loss.Furthermore,it effectively enhances the embedding capacity by fully utilizing the correlation between adjacent pixels while ensuring the security of the image.
