Research On Reversible Data Hiding For High Dynamic Range Images

Computers and social networks have been rapidly developed today. Digital life and office gradually become normalized. Enjoying the convenience of the Internet transmission and sharing, businesses and individuals are facing the risk that the listeners steal their user information even sensitive information. As an important aspect of information security, reversible data hiding have important applications in the military, legal evidence and medical imaging fields. This technology hides data into multimedia files and makes information invisible. In addition, the original files and data can be completely restored at the receiving side.Digital images are common files for reversible data hiding technique. With the development of multimedia technology, a new image format, high dynamic range image can record the true scene. Different with conventional low dynamic range image, the data structure of high dynamic range images is complex, because it not only records the color information as well as luminance information. Hence it is different to apply reversible data hiding techniques on such images. Based on the system research on information hiding technologies, the paper proposes a scheme for high dynamic images and does related experiments. The detail contents and innovation points are listed as following:1. High dynamic range images has sufficient color and lighting information, but not easy to deal with its data. In this paper, the image is encoded by RGBE which can convert floating-point data into integer data. The method not only reduces storage space and makes the image easier to process.2. The key of the reversible data hiding is the accuracy of prediction model. Take advantages of relation among the channel, the paper divides image into edge and smooth region and adopted different prediction model for them to enhance the efficiency of data hiding.3. The degree of distortion for each pixel is different. Therefore, considering the prediction accuracy, the possibility of the overflow problems and the distortion caused by light, this paper constructs a sorting strategy to ensure that data is always hidden in pixels with small distortion to further reduce the distortion.4. Before data embedding process, based on the amount of data to be embedded, the thresholds in this algorithm are initialized to optimize this algorithm and improve the efficiency of the algorithm.Experiments show that the proposed reversible data hiding algorithm on high dynamic range images is superior to the traditional ones. And it achieves a tradeoff between embedding capacity and distortion. This algorithm enriched the theoretical research on the high dynamic range image fields, and has a contribution for data security.