Reversible Data Hiding Algorithm With High Visual Quality

With the rapid development of the Internet and information technology,people's lives are becoming more and more convenient.However,this also brings a series of problems,such as copyright protection,transmission and distribution of digital content in recent years.Therefore,information security research has aroused widespread concern.Reversible data hiding technology is one of the effective means to solve information security.In addition to ensuring the confidentiality and extractability of embedding data,the original carrier also needs to be restored completely without damage.Reversible data hiding is a special information hiding technology.Compared with traditional information hiding technologies(such as digital watermarking and steganography),it has its own advantages.Digital watermarking focuses on how to reliably extract data from potentially distorted marking media.This technique takes robustness as a top priority,but pays little attention to the recovery of the original carrier.The steganography focuses on hiding the secret information into the original carrier in an undetectable way.It pays more attention to the imperceptibility of hidden data and embedding operations,i.e.not only the human eye cannot perceive the secret data,but also the potential analyzer can't detect the embedding operation.Compared with these two data hiding technologies,reversible data hiding technology pays more attention to the non-destructive recovery of the original carrier and secret information,so it is applicable to fields that do not allow modification of the original carrier such as medical,military,and court.This dissertation mainly studies reversible data hiding algorithm with high visual quality.The main research work and innovation results of the thesis can be summarized as the following points.1.A reversible data hiding algorithm based on multi-texture is proposed.In order to improve the contrast of the medical image and achieve a higher embedding capacity,the background of the medical image is separated from the foreground using the texture of the medical image,and the contrast of the entire image is enhanced by a contrast enhancement algorithm.In order to improve the contrast of the key parts of the medical image further,the method similar to the histogram equalization effect is adopted embed the data for the foreground,thereby the foreground contrast of the image is enhanced further.The background portion does not affect the use of the image,and the grayscale value is single,so that a large amount of information is embedded using the portion.The algorithm proposed is not only better than the existing algorithm in contrast enhancement,but also can reach a high embedding capacity.2.A reversible data hiding with different embedding capacity based on optimal embedding strategy selection and image quality assessment criteria is proposed.Since the existing algorithm cannot take care the high embedding capacity and the low embedding capacity,a reversible data hiding with different embedding capacity based on optimal embedding strategy selection and image quality assessment criteria is proposed in this dissertation.Since the results of the traditional Peak Signal-to-Noise Ratio(PSNR)are not consistent with the subjective feeling of human eye sometimes,another two image quality evaluation indicators that are more consistent with human eye vision are employed in this algorithm.Since the emphasis of each image quality assessment criteria is different,and the causes of distortion are different under the high embedding capacity and the low embedding capacity,optional embedding strategy based on IQA criterion with different EC is proposed in this dissertation.Experiments show that the proposed algorithm is better than the traditional reversible data hiding algorithm in both high embedding capacity and the low embedding capacity.3.A reversible data hiding with skin tone smoothing is proposed.Almost all face images in daily life have been beautified,this is because beautifying the face part can improve the visual quality of the image better.Therefore,this dissertation proposed a reversible data hiding algorithm with skin tone smoothing,which aims to achieve skin beautification and information embedding at the same time.The algorithm first embeds the secret data into the original image,and then performs a skin tone smoothing operation on the marked image.The decoding first needs to restore the confidential image before extracting the data,and finally extracts the secret data to restore the original image.Experiments show that the algorithm can achieve better skin tone smoothing effect after data embedding,and the original image can be restored losslessly from the beautified image.