Research On Image Steganography Based On Modified Direction Encoding

With the popularity of the Internet and the rapid development of information processing technology and communication means,it brings convenience to communication but also poses new challenges for information security.Covert communication is mainly used to transmit confidential information such as military intelligence in the network,and it is a key aspect of information security.Steganography plays an important role in covert communication and intelligence work because of its unique imperceptibility and anti-detection characteristics.At present,the research on steganography mainly focuses on improving the embedding capacity,image qualityand resistance to the steganalysis and other performance indicators.In this work,based on the application of steganography in covert communication technology,an improved steganography algorithm based on modified direction encoding is proposed.The work done in this work mainly has the following three aspects:(1)The overall scheme design of covert communication image steganographyBased on the analysis that covert communication requires reliable transmission of operational information,the technical requirements of information hiding for covert communication are summarized:high quality,medium embedding capacity and safety.By summarizing and comparing the advantages and disadvantages of various spatial domain steganography algorithms,the overall scheme satisfying covert communication steganography is determined.The adopted EMD(Exploiting Modification Direction)encoding steganography guarantees image quality and pixel bit plane segmentation technology is used to improve the embedding capacity.The design about the minimum numerical substitution method and the data preprocessing method improves the anti-aggression of the algorithm.(2)Research on base conversion method of the big data sequenceThe large number encoding serial conversion is the key to steganography algorithm based on the modified direction,but the existing technology has little research on it.For the first time,we systematically compared the limitations of the current various types about base conversions,and we proposed the idea of segmentation vertical long division to solve the problem of overflow and conversion omission;the remainder residual error among segments in each iteration is eliminated by the segment remainder transfer function.The experimental results show that this method can correctly achieve the large-scale conversion of arbitrary bytes,and the base value can be any number,which breaks the limitation of the built-in conversion and realizes the versatility of the conversion.(3)Research on image steganography algorithm based on modified direction encodingUnder the premise of maintaining the image quality of the existing EMD method,an improved image steganography algorithm based on EMD is proposed.The work mainly made the following improvements:Improve embedding capacity and security with base conversion;the work overcomes the shortcomings of the fixed bit-segmentation method in the traditional bit-plane segmentation method and doubles the embedding unit.For the two low planes,the minimum number substitution method is used to embed the converted new cryptographic secret number,which improves the embedding capacity of the algorithm.For the remaining 4 bit planes,the EMD method is used to embed information according to scrambling coding,the proposed algorithm ensures higher PSNR and SSIM of the hidden image,and improves the anti-aggression of the algorithm.Experiments on the SIPI standard library show that the embedding capacity of the proposed image steganography algorithm based on modified direction encoding is increased by approximately 15%when the average PSNR is kept at around 41 dB comparing with other steganography algorithms.The maximum embedding capacity of the algorithm can reach 810000 bits when the PSNR value is maintained at about 37 dB.