Design And Implementation Of High-capacity Data Hiding Algorithm In Encrypted Image Domain

With the rapid development of multimedia and communication technology and the prevalence of mobile devices, digital images, as the most intuitive and vivid cover, play a more important role in people’s daily life. Digital images are easily hacked, copied and tampered, so the security of digital images becomes an imminent and urgent problem. Encryption is a popular way to solve this problem. Encryption turns the images to incomprehensible data to the unauthorized users. Data hiding exploits the redundancy of the cover and people’s insensitivity to embed the information to the cover in some way. The data extraction is the inverse process of the data embedding. It is used to authenticate identity, annotate the content and protect the copyright, etc.In many occasions, the encryption and data hiding are processed by two people. To protect the confidentiality and the privacy of the image, the content owner encrypts the image and sends it to the data hider to embed data. As a consequence, data hiding in encrypted images becomes more important. It has the potential to be widely applied in medical diagnosis, cloud storage and remote medical treatment. In this thesis, we propose high capacity data hiding schemes in encrypted images.Our main work is listed as below:① By researching the mainstream encryption and data hiding algorithms, systematic knowledge about cryptography is obtained. And then the existing data hiding algorithms in encrypted images are deeply studied to master their merits and demerits, which can lay solid foundation for further improvement and upgrade.②The compressive sensing(CS) theory is thoroughly studied in the aspects from signal sparsification, designing the measurement matrix, measuring the sparse signal and signal recovery. And then compressive sensing is utilized to design the data hiding scheme in encrypted image domain, which is a new direction for CS application. At the same time, the scheme can embed more data and is quite robust.③Based on the correlation of the neighboring pixels, a new high capacity reversible data hiding in encrypted image algorithm is proposed. We flip the MSB of the pixels based on the embedding data, and extract data and recover original image by calculating the difference between the pixel value of the decrypted image and the estimate value. The algorithm is reversible and of high capacity with the embedding rate around 0.5 bpp. It is rare in reversible data hiding algorithms in encrypted images. The extracted data is free of any error and the PSNR of the directly decrypted image is high.