The Study Of Image Scrambling Algorithm Based On Doppler And Evaluation Method

Since its advantage of visual and information-rich,the images are more and more popular with people of all ages, and are more and more relied as a medium for exchange of information in the rapid development of the Internet today. However, in the process of image storage and transmission over the network,it is vulnerable to the illegal destruction of the attacker, especially the images which are relation to national security, military secrets, and individual privacy, once being illegally intercepted, serious damage will bring to the interests of States and individuals. Therefore, how to ensure the safety and effectiveness of these images is especially important. Image scrambling technology is an important aspect of information encryption, its main purpose is to randomize the image information, enhanced information hiding, and able to withstand attack. Based on the extensive literature study, the paper found that there are still some deficiencies, such as: some has size requirement, some has recovery cycle, some has small key space, some has computational complexity. Therefore, for above insufficiencies, the paper proposed a new image scrambling algorithm; There are a lot of degree evaluation methods, and most of them are based on the spatial territory, conversely, there are few in the wavelet territory, and some are short on the change of position and gray information, and some results of the evaluation are inconsistent with the subjective feelings. Therefore, The paper also propose a new evaluation method of image scrambling algorithm. The content of this article is divided into the following two aspects: Image scrambling algorithm based on Doppler and Wavelet transform, The image scrambling is applied to the wavelet territory, its position and gray values while scrambling. Firstly, the image is divided into blocks, then the block is processed by Wavelet transform, lastly use Doppler carry pixel position and values scrambling on Wavelet coefficients. Experimental results show that the algorithm can achieve better scrambling effect than existing algorithms, change the statistical property of the image, and have great improvement in visual and resisting attacks, and can be applied to image scrambling efficiently.Image scrambling degree evaluation on distance and texture of wavelet territory. Firstly, the images before and after scrambling are transformed by lifting wavelet respectively, Secondly, calculate the Manhattan Distance for corresponding wavelet coefficient, and then generate the gray level co-occurrence matrix based on itshigh-frequency coefficients, extract the texture features,and finally scrambling degree is obtained by the change of distance and texture. The method is based on the wavelet territory, regardless of neighborhood relations, which is more accurate than methods in spatial domain. Experimental results show that the evaluation effects correlate well with subjective assessment, and measure the extent of scrambling, and reflect the relationship between the frequency and extent, which all reflect the effectiveness and practicality of the algorithm.