Research On Coverless Text Information Hiding Based On Frequent Words In Text Sets

Information security involves a country's political,military,economic and other aspects of security.Information hiding technology,which has been widely used in secret communications and copyright protection,has been becoming a hot direction in the field of information security research.Traditional information hiding methods are mostly carried out by embedding information in the carriers,which is bound to change the statistical distribution of the carriers.An attacker can know the presence of hidden information or even extract it by analyzing the feature changes.In 2015,coverless information hiding,which proposed by experts of China Information Hiding Workgroup,is a good idea to solve this problem.Coverless information hiding technology can directly generate or choose stego-vectors from the secret message.The carriers including the secret information can be sent to the receiver without any modification of it.Coverless information hiding technology does not embed information into the carrier,so it does not change the statistical distribution of the carrier characteristics,and could resist the existing steganalysis detection method based on abnormal detection.Coverless information hiding was positioned as the second generation information hiding technology at the 13^th China Information Hiding Workgroup held in 2016,and has become one hot direction in the filed of information hiding research at present.In this thesis,some text coverless information hiding methods were proposed.These methods could be implementd in the English texts big data environment.The main research work and results are as follows.?1?By analyzing the structure and characteristics of OOX documents,we proposed a novel information hiding method by using the custom component to hide the secret message.Since the custom components are not referenced in the main document part of the OOX document,there is no visual differencethe for the carrier,and the secret message has a better concealment.At the same time,the custom components exist undependently of the main document parts because they are not referrenced,so the information hiding process does not cause any changes in the presented contents.In this sense,the proposed method is a coverless information hiding method based on OOX documents.?2?With the support of the text big data,we analyzed the distribution of word frequencies in a text set,and proposed a novle text coverless information hiding method based on the Word Rank Map by firstly defining the Word Rank Map for the text set,a text and a word.Firstly,we transformed the secret message into the frequent words by implementing the transform protocol agreed upon in advance by communication parties.Secondly,we chose the rank labels for the transformed secret message by implementing the rank label location protocol agreed upon in advance by communication parties.Thirdly,normal texts containing the transformed secret message and the corresponding labels will be retrieved from the text big data,and they coubld be chosen as stego-texts.Finally,the chosen texts will be sent to the receiver without any modification,and the converless information hiding is achieved.?3?Inspired by the Simhash algorithm used by Google for duplicated Web pages detecting,we defined the frequent words hash signature of a text in a text set,and so a text will be mapped into a bits string accociated with the frequent words.On this basis,we defined the frequent word distance for a text in a text set,and chose corresponding rank labels for the transformed secret message by using this distance and the rank label location protocol.Finally,normal texts containing the transformed secret message and the corresponding labels will be retrieved from the word rank maps of the frequent words,and they coubld be chosen as stego-texts.So the converless information hiding is achieved.?4?In order to easily transform the secret message into the frequent words in a text set,we designed two kinds of word transform protocols and word rank label location protocols by combining the word rank maps and the frequent word hash.By implementing these protocols,normal texts containing the transformed secret message and the corresponding rank labels will be retrieved from the text big data,and they could be chosen as stego-texts without any modification.So the passing of the secret message could be achieved by sending the stego-texts.This thesis provides a new idea for the research of text coverless information hiding.The research results could be easilly extended to the Chinese texts,and will be used in the field of information security such as secret communications.