Study Of Mathematical Theory And Key Technology For Digital Image Steganography

With the rapid development of the information hiding technology and network, information transmission becomes more and more convenient on computer network. However, bring conveniency to communication, it has brought to forefront security concerns with the use of digital images. As an important method of information security transmission, information hiding technology has attracted great concern. Steganography, as the main embranchment of information hiding, hiding the secret information into innocent cover and transmits on the public channels, is more imperceptible. Steganography mainly focuses on embedding capacity, embedding distortion and security. This paper is mainly researching on steganography techniques of images.A novel data hiding method using pixel-value difference and modulus function for color image is proposed. The proposed method has fully taken into account the correlation of the R, G and B plane of a color image and human eyes are sensitive to the change in color images. The amount of information embedded the R plane and the B plane determined by the difference of the corresponding pixel value between the G plane and the median of G pixel value in each pixel block.Furthermore, two sophisticated pixel value adjustment processes are provided to maintain the division consistency and to solve underflow and overflow problems.The most importance is that the secret data are completely extracted through the mathematical theoretical proof.A steganographic scheme by fully exploiting modification directions was proposed by Kieu and Chang, and we point out that it has a loophole. Kier-Chang scheme requires about 524Kbytes(256 × 256 × 8 = 524.288Kb) to store the embedding matrix, because they use the search matrix method to finish embedding the secret data, and the original EMD method considers no image characteristic and embeds digits in the same base into the whole cover image; therefore, each pixel pair embeds equal amount of data. A data hiding scheme using pixel value differencing and improving exploiting modification directions is proposed. A cover image is first mapped into a 1D pixels sequence by Hilbert filling curve and then divided into non-overlapping embedding units with two consecutive pixels. Our method exploits pixel value differences(PVD) to estimate the base of digits to be embedded into pixel pairs. The proposed method modified the EMD method so that the digits in any bases can be embedded according to the local complexity of the cover image. Our method does not require any memory space to store the embedding matrix.A side-match predictions applied in data hiding method that adaptively embeds data into pixel value by using improved exploiting modification direction(EMD)method is proposed. And we take into consideration the correlation between the pixel and its surrounding pixels on each 3 × 3 block unlike the traditional EMD which concealed an equal amount of data into each pixel and cause equal degree of distortion. The method exploits pixel value differences(PVD) to estimate the based of digits to be embedded into pixel value. In Lee et al.s interpolation method,due to being only an approximation, an interpolated image will always lose some quality each time interpolation is performed. In view of the problems existing in the two above method, the proposed method adopt the concept of PVD and propose a data hiding method based on EMD with the consideration of human visual system(HVS). And the proposed scheme applied to efficiently improve the number of the embedding bits with low computational complexity and good image quality.A high-capacity data hiding method based on the index function and PVD is presented. The proposed scheme divides a cover image into non-overlapping subblocks, and the basis pixel is decided by the improved index function. Difference values with the other three pixel value in each sub-block are referenced to decide the number of embedding secret bits with the range table. In addition, the proposed method effectively controls the distortion between the stego image and the original image by using remainder function. In the extracting process, the proposed method subtracts the qualified basis pixel value from each stego images to obtain the secret data.