A Steganography Research Based On Visual Perception Mechanism

Information hiding is a new technology for the information security, and has gradually become a booming research. Digital steganography is a key branch of information hiding, which plays an important role in daily life. Steganography capacity and visual imperceptibility are the most important indexes to measure a steganography algorithm. We must make a tradeoff between the two important indicators owing to the restricted and contradictory of them. A steganography research based on visual perception mechanism can make a balance between the two important indicators. We focus on the algorithms, and then make some achievements.The first, most existing image quality objective metrics had some inconsistent with human’s subjective judgment. We reasonably analyze the advantages and disadvantages of objective metrics, and design the image quality multiple index evaluation scheme. The multiple index evaluation is of great significance owing to the lack of unified image quality metrics, and then provides a research support to a steganography research based on visual perception mechanism.The second, visual attention is a very important characteristic in the human perception system. Most steganography algorithms based on the visual perception mechanism just considered the non attentional factors, such as brightness, contrast and masking effect, and neglected the importance of the visual attention for the steganography algorithm. A new gray image steganography algorithm based on visual attention and local complexity was proposed. The algorithm is to analyze the local complexity of the image via standard deviation, introduce the Itti visual attention model in the more complex areas and then quantitatively depict attention characteristic using visual entropy, hide the message according to the visual attention and local complexity finally. The final experimental results were carried out showing that the proposed algorithm still maintains good imperceptibility after embedding large-capacity information, and can resist the histogram contrast steganalysis.The Third, most existing steganography algorithms for color images were designed by directly applying steganography for gray-scale images to color channel without employing color visual perception mechanism. A large-capacity blind steganography algorithm for color images based on visual perception mechanism in YUV space was proposed. The main idea is to analyze the local complexity of the cover image via standard deviation on the Y component, split the message into two parts using the color visual perception mechanism, and then hide them in Y and V components respectively. Being less flexible, the U component is simply used as an information indicator. The final experimental results were carried out showing that the proposed algorithm still maintains good imperceptibility compared with other methods after embedding large-capacity information, and can resist both the histogram contrast and RS steganalysis.