The Application Of Compressive Sensing In Information Hiding

Compressive sensing is a breakthrough technology of signal sampling in recent years. It can sample and compress data at the same time and thus can avoid storing large amounts of data by traditional sampling algorithm. Compressive sensing has been successfully applied in face recognition, image denoising, digital forensics, etc. Based on the studing of compressive sensing theory, measurement matrix building and reconstruction algorithms, this paper uses compressive sensing in information hiding as following:First, this paper compares the qualities of reconstructed images under different measurement values and different image reconstruction algorithms and concludes the relation among reconstruction quality, measurement value and reconstruction algorithm. By analysing some existing information hiding method based on compressive sensing, this paper develops new methods.Secondly, this paper proposes an Image encrypting algorithm based on compressed sensing. Taking advantage of measurement matrix can scrambling as well as compressing images, this algorithm chooses Gaussian distribution matrix as measurement matrix and multiply with image to encrypt it.Thirdly, this paper proposes an information hiding algorithm based on compressed sensing. This algorithm puts secret image and part of original image multiply by measurement matrix, then subtraction each other and embeds to original image’s middle band in the DCT domain. This algorithm is of high security and has enough robustness.Fourthly, this paper presents a local tamper detection algorithm based on compressed sensing. By using measurement matrix multiply with protecting areas of image and watermark data, then embeds the subtracted each other’s data. This algorithm causes the algorithm to be more sensitivity with tamper.