Image Scrambling Degree Evaluation Based On The Fusion Of Neighborhood And Texture Degree

Image scrambling is one of the image encryption techniques with a wide application in data hiding and watermarking protection. Nowadays, people pay more and more attention to the encryption function of image scrambling. Therefore, the evaluation to scrambling degree is of great importance. There exist two methods of scrambling degree evaluation. First, subjective eye evaluation is intuitive and accurate unless the image is beyond our judgement. Another method is computational evaluation, which results in scrambling degree algorithms. A perfect evaluation algorithm would usually provide the same result as the subjective eye evaluation.Existing scrambling degree algorithms are based on pixel change, displacement change, neighborhood, pixel uniformity distribution and etc. However, these algorithm cannot catch well the degree of image scrambling.Basd on the analysis of existing evaluation standards, our thesis focuses on the evaluation algorithms based on neiborhood and proposes two novel evaluation algorithms to calculate the image scrambling degree. Our contributions are below:(1) Classify existing scrambling degree evaluation standards.(2) Analyze the neighborhood-based evaluation algorithms in details, and summarize the disadvantages of the algorithms.(3) Propose a texture-based evaluation algorithm to calculate image scrambling degree. Moreover, combine the neighborhood-based and texture-based standared to design a novel effective evaluation algorithm(4) Experiment with real image data and draw conclusions.