Research And Implementation Of A Semi-creative Coverless Information Hiding Algorithm

In the information age,how to improve the security of information storage and transmission has become an urgent problem to be solved and a hotspot of academic research.In order to improve the security of information transmission,the information hiding technology,as one of the methods to realize covert communication,has developed rapidly in the past twenty years.Information hiding technology mainly includes steganography,digital watermarking technology,etc.However,the steganography analysis technology,which develops along with steganography technology,also gets the development of antagonism and becomes a detection method to detect whether the carrier hides secret messages.With steganographic methods based on embedded methods becoming a bottleneck in recent years,in order to improve the security of steganography,the coverless information hiding,as a method of information hiding without modifying the carrier,has become a hot research hotspot in the field of information hiding.In order to improve the data embedding capacity of coverless information hiding algorithm,scholars have done a lot of research work to promote its practical research.On the basis of studying the coverless information hiding algorithm,this paper proposes a semi-structured coverless information hiding algorithm combined with the advantages of convolutional neural networks to improve the capacity of data embedding and the efficiency of communication.The special idea is to analyze and extract the user's behavioral habits by researching several social platforms,the small icons are spliced together based on the construction rules from behavioral habits and the mapping relationship established between small icon library and binary random numbers,which combined with the text description of construction rules for rationality of content and logic to implement the graphical representation of secret message and achieve the transmission of secret message by sending those spliced image through the social platform.The algorithm proposed in this paper reduces the difficulty of building image library at the sending end,but makes higher demand for the correct classification of small icons at the receiving end.For this reason,the convolutional neural network is introduced to train and classify icons in the icon library,which could realize the correct classification of small icons.In order to improve the robustness of the algorithm,some icons processed by various attack methods are deliberately added as interference samples in the training set when using convolutional neural network for transfer learning.Due to the presence of various images on social platforms,and users of social platform might modify the images and republish them on the social platform,therefore,the paper also designs a watermarking algorithm,which adds dark watermark to the secret image,and verifies the authenticity and integrity of the secret image through detecting the watermark information.To verify the validity of the algorithm,this paper also build a covert communication system using the algorithm with corresponding programming language to analyze the performance index such as robustness and capacity of the algorithm.According to the experimental results,the algorithm proposed in this paper has a good anti-attack ability and improves the hiding capacity,which can be used to achieve convert communication.