Research And Design Of Data Hiding Algorithm In Encrypted Domain Based On Compressive Sensing

With the rapid growth of modern digital communication and electronic data exchange,we have entered the era of informatization that combines "Internet +" and big data.Digital images have been widely used in different fields such as medicine,commerce,and military due to their vivid and intuitive features.When digital information is transmitted in insecure channels,there will be security issues such as privacy violation,malicious tampering,and illegal copying.Therefore,it is necessary to provide appropriate security for images to prevent unauthorized persons from accessing important information.Compressive sensing(CS)is a new signal sampling theory,which breaks the constraints of Nyquist’s sampling theorem.It can effectively reconstruct signals with high probability by a small number of linear and non-adaptive measurements.The CS framework can also be viewed as a symmetric encryption system for protecting image content.The information hiding technology uses the image redundancy to hide the secret information,which solves the problems of image copyright protection and integrity authentication.Therefore,it is of great significance to combine CS with image information hiding technology to carry out in-depth research on information hiding in encrypted domain.The work of this thesis completed mainly includes:(1)A double watermarking algorithm for medical images in encrypted domain based on CS is proposed.Two watermarks are directly embedded on the CS measurements of the cover data(the encrypted medical images): One is the authentication watermark,which is used to authenticate the origin of the data;the other is the annotation watermark,which is used to record various information during usage of the medical image.According to their different permissions,data receiver can decrypt the original image and extract different watermark information.The proposed scheme completely protects the security of the carrier data and watermark data,and ensures that users with different authorities can obtain different watermark information.Experimental results verify the performance of the proposed scheme.(2)Based on the relevant knowledge of tensors,a separable large-capacity data hiding scheme based on Tucker decomposition in encrypted color images is proposed.This scheme makes full use of the characteristic images generated by Tucker decomposition for image encryption and data embedding.At the same time,it provides a new idea for data hiding of color images.This scheme realizes the separation of data extraction and image restoration.It has a high embedding ability,and can resist a certain degree of attack.