Research On Multi-carrier Secret Map Sharing Combined With EMD Steganography

Image sharing is the method that shares the secret image into multiple distributed images,and these distributed images can be used to reconstruct the secret image completely.Conventional image sharing based on Lagrange interpolation usually is designed very complex to guarantee secret image and distributed images in equal size or free image expansion.Different from conventional image sharing,multi secret image sharing strategy encrypts multiple secret images into meaningless images that are transmitted in the public channel based on XOR operation.Meanwhile the sharing images are held by different participants.And only part or all the sharing images can be used to correctly reconstruct the secret images.The problem is that in the transmission of the public channel,the sharing images are easy to be destroyed by a malicious attacker.Thus the possibility of reconstructing secret images is reduced.While multi-carrier secret image sharing method provides a simple and effective approach for image sharing.It usually uses(N,N)-threhold to hide one secret image into multiple same size carrier images and when only all distributed images are collected,the secret image can be completely reconstructed.However,in conventional multi-carrier secret image sharing schemes based on error diffusion and recovery function,the error diffusion method always leads to low visual quality and the recovery function is only designed for specific image resolution.Moreover,it can only provide limited security by simple Arnold scrambling or XOR encryption.To address these problems,the followings are the finished works:1)A multi-carrier secret image sharing scheme combined with EMD-cl embedding was proposed.In the proposed algorithm,first the double hash:MD5 and SHA-1 values were used to randomly generate secret image and user keys related scrambling variables which were used to change every carrier image pixel position by 2D bi-scale rectangular mapping.Second vector composed of same position pixels in scrambled carrier images were randomly allocated weights of EMD-cl by extended Josephus mapping which were related to start position,count termination value sequence,count gap sequence and count direction sequence.Finally EMD-cl was used to embed secret image pixels into multiple cover images.By comparing with conventional methods,the visual qualities of all stego carrier images are enhanced and the proposed scheme needn't design any recovery function and can be easily applied into different resolution or gray scale images.There are improved security that carrier image pixel mapping relationship and the basis vector in EMD-cl embedding method were bounded with the MD5 value,SHA-1 value and the user keys.So only the correct user keys,MD5 and SHA-1 values can recover the secret image and both the MD5 and SHA-1 values of key related variables sharing information are public by the third trusted party to provide some authentication ability.The proposed method has low complexity,high security,universality and provides some authentication ability.So it is superior to the conventional multi-carrier secret image sharing scheme based on the error diffusion and recovery function in the overall performance.2)In order to further decrease visual quality influence of EMD-cl for distributed carriers,increase authentication and recovery abilities for extracted secret image,free of the keys and variables managed by the third trusted party,and avoid poor universality and low flexibility by binding secret image type and carrier number,a multi-carrier secret image sharing scheme combined with EMD-2l embedding and Hamming code was proposed.In sharing phase:firstly,the secret image was changed into the binary secret hamming checkout sequence by adding the hamming checkout bits;Secondly,all participants' keys were used to generate the main key which was used to generate the random gray image with the same size of cover image;Thirdly,the binary secret hamming checkout sequence was scrambled by the logistic chaotic mapping where the variables of logistic chaotic mapping was also generated by the main key.Finally,the scrambled hamming checkout sequence were grouped based on the number of the distributed carrier images and embedded into the corresponding carrier pixel group one by one by EMD-2l to form the distributed carriers.In recovering phase:firstly,the random gray image and the variables of logistic chaotic mapping were generated by the main key.Secondly,the binary secret hamming checkout bits sequence were grouped based on the number of the distributed carrier images.The scrambled binary secret hamming checkout bits sequence was extracted by EMD-2l and recovered by logistic chaotic mapping.Finally,the recovered binary secret hamming checkout bits sequence was divided into the secret image bits sequence and hamming checkout bits sequence.Meanwhile both were grouped by the number of the distributed carrier images.The secret bit information of every group was detected error and corrected error by hamming code.The experiment results show,compared with conventional multi-carrier secret image sharing scheme,the proposed method can provide high visual qualities of all stego carrier images and embed arbitrary resolution and gray level secret image.It has low complexity,universality and provides some authentication and error-correcting ability.Compared with the proposed multi-carrier secret image sharing scheme combined with EMD-cl embedding,the proposed method provides higher visual qualities of all stego carrier images and has certain error detection and error correction ability for the recovered secret information.