Reversible Data Hiding In Encrypted Images Based On Ldpc Code

With the development of cloud computing and other technologies,people's lives have been greatly enriched,and the communication between people has become increasingly close.With the use of applications such as WeChat,people have developed from text communication to multi-media communication such as voice,image and video.Although ways of communication and lifestyle have been greatly improved,there are also some security risks.The most important one is the leakage of privacy.In the past,because of the level of development,people paid more attention to protecting the security of text information.However,with the development of technology and the application of multimedia technology,the security of multimedia information such as images has been proposed.In the field of military and medical,the security of multimedia information is particularly important.Nowadays,face recognition,iris recognition and other technologies are widely used,which bring greater challenges to the protection of image information.How to protect image information while hiding secret information has attracted many researchers.In order to improve the security of image and secret information,we propose two reversible information hiding algorithms for encrypted image based on LDPC code.These two methods can not only protect the information security of the original image,but also hide secret information to achieve the result of information protection.In the first method,we first encrypt the original image by using binary XOR and block scrambling.Then the data hider divides the encrypted image blocks into smooth region and complex region according to a threshold and collects the least significant bits(LSB)of the pixels in the smooth region.Then these bits are compressed by LDPC matrix,and the additional bits are embedded into the compressed room to generate a new bit stream.In this way,if the receiver only has the encryption key,he can decrypt the image directly but can not extract the additional information.If the receiver only has the data hiding key,he can extract the additional information directly.If the receiver has both the encryption key and the data hiding key,he can extract the additional information correctly and recover the original image without any error.In the second method,we encrypt the original image according to stream-cipher and block permutation to generate the encrypted image.After getting the encrypted image,the data hider divides the encrypted blocks into smooth region and complex region according to the threshold.Unlike with the first method,we directly replace the most significant bits(MSB)of some pixels in the smooth region with additional information to complete the first step.Then the least significant bits(LSB)of other pixels are embedded again by LDPC matrix compression to complete the second step.In this way,the receiver can decrypt the marked image directly if he only has the encryption key.If he only the data hiding key,he can extract the additional information directly without any error.If he has both encryption key and data hiding key,he can not only extract additional information accurately,but also recover the original image.Our schemes improve the embedding capacity and the image quality of decrypted image.