Research On Evaluation Of Information Hiding Based Digital Image

With the rapid development of Internet, information security has been concerned widely for the increasinglyfrequent transmission and exchange of multimedia data. Information hiding is a new kind of information securitytechnology. A lot of information hiding algorithms have been proposed for years. However, for the securityassessment of information hiding system, there has never been a comprehensive evaluation model. Therefore,evaluation of information hiding plays an important role on the establishment of evaluation criteria and theimprovement of information hiding algorithms.For the problems in the evaluation of information hiding, the consistency of image quality of subjective andobjective evaluation has been studied in this paper. The research mainly covers following sections:(1) By analyzing the performance impact of different matching orders between secretary messages andcarrier pixels on the algorithm, this paper proposes a three level score evaluating strategy based on the pair-wiseLSB (Least Significant Bit substitution) matching method proposed by Mielikainen. According to the newevaluating strategy, a particle swarm optimization algorithm is applied to search for an optimal solution among allsuch permutation orders. Afterward a novel steganographic algorithm is presented with respect to Mielikainen’smethod. The experimental results indicate that the proposed method reduces the number of changed cover pixelvalues after embedding the secret messages with the same amount of secret information. Therefore, the stegoimage achieves higher visual perceptual quality than Mielikainen’s method.(2) Due to the bad performance of the imperceptibility evaluation of information hiding for subjectiveevaluation, a new method based on visual masking is proposed. It combines the linear relationship between theperceptual quality of image and its Mean Square Error, and the perceptual properties of texture and luminance,with the aim of improving the perceptual quality. In this method, the linear expression between the perceptualquality and Mean Square Error has been given, among which the luminance and the gradient can be used tocalculate the luminance and the texture weight coefficients respectively. By weighting the expression linearly, aglobal performance of imperceptibility of the stego image is obtained. Experimental results demonstrate that theproposal outperforms the conventional Peak Signal Noise Ratio in higher efficiency and realizing the consistencyof image quality of subjective and objective evaluation.