Research On Coverless Information Hiding Method Based On Image Content Mapping

In recent years,as the public attaches more and more importance to information protection,information security has become a key concern for everyone.Information hiding technology has been widely used in anti-counterfeiting and ensuring the secure transmission of secret information.The traditional information hiding technology destroys the naturalness of the cover because the secret information is embedded in the cover,which is easily detected by the steganalysis algorithm.In response to this problem,"non-embedded" technology came into being,that is,coverless information hiding technology.The coverless information hiding technology does not embed secret information in the cover,but uses the information of the cover itself to represent the secret information.Thereby it can achieve secure transmission.Therefore,the coverless information hiding technology has important practical significance.However,in the development process of coverless information hiding technology,problems such as low hidden capacity and poor image quality have occurred.After in-depth understanding of multiple solutions for coverless information hiding,this thesis proposes a coverless information hiding method based on image content mapping.The specific research results are as follows:(1)A coverless information hiding technology based on image block matching is proposed.In this method,the stego image is synthesized by retrieving similar image blocks in the image database according to secret information.Each image block represents 1-bit secret information and the hidden capacity can reach 30000 bits.Experimental results show that this method is superior to other coverless information hiding methods in terms of capacity and image visual quality.(2)A coverless information hiding technology based on star GAN and face image content mapping is proposed.This method takes advantage of the feature that star GAN can realize the characteristics of multi-domain conversion of the image.Driven by secret information,a high-quality stego image is generated by star GAN to realize the secret transmission of information.Experimental results show that this method improves the hidden capacity,and also avoids the problem that the GAN network cannot completely restore the secret information,and maintains the high quality of the image.