Research On Visual Cryptography Scheme For Cheating Prevention

Visual cryptography is directly decrypted by human visual system, and it is similar to "one time one pad" cryptographic system in intensity of security. There are cheating problems in visual cryptography as in other secret sharing schemes. Although most visual cryptography schemes assume that participants are honest and do not cheat when decrypting the secret image. But they are not all honest participant which would generate fake shares to cheat others. So it has grand theoretical meaning and applied cost to design cheater preventable visual cryptography schemes.In general, the basic matrixes are constructed by linking basis matrixes of some threshold visual cryptography schemes, and let them meet the requirements, then generate shares based on them. In this way, the pixel expansion is bigger and contrast is smaller.Firstly, the paper studies the types of fraud and anti-fraud method deeply and describes the cheating methods in visual cryptography scheme. The participants can cheat together in visual cryptography scheme with public verification image, concrete cheating method is given and simulation is implemented. For these schemes can not be implemented in two participants, multi-secrets sharing schemes are applied in cheater detectable visual cryptography scheme based on pixel coding. The secret image and the verification images can be revealed by stacking shares in different ways.Secondly, we analysis the verification matrixes in general transformation for cheating prevention scheme(GTCP) with private verification image, verify the construction defects and it leads that participant can cheat in the scheme, then show the concrete cheating method. By improving the verification matrixes, a visual cryptography scheme with cheating prevention is proposed, and proof the existence of the matrixes, compared with GTCP scheme enhance the security.Finally, a new visual cryptography scheme for cheating preventin with private verification image is designed and implemented by studying the basis matrixes. In order to avoid cheating behaviors, and the parameter be near optimal both contrast and pixel expansion, we present the concrete requirements for the matrices and introduce the method of constructing the matrices. By combining the idea of having private verification images and based on the verification matrixes construct the basis matrixes, then generate shares based on them, the secret image and the private verification images can be shared. Compared with the other anti-fraud schemes, ours can effectively avoid the cheating problem and reduce the matrixes connection number, the pixel expansion is reduced and the contrast is increased. The simulation results show that the secret image and the verification images is recoverd better.