Font Size: a A A

Information Hiding Based On Intelligent Generation Of Tang Poetry

Posted on:2019-03-08Degree:MasterType:Thesis
Country:ChinaCandidate:B B LiFull Text:PDF
GTID:2438330548473739Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the rapid development of computer science and network technology,the era of big data and information has come.Network security and information security are the important research fields in information era of science and technology,which have risen to the height of national strategic security.At the same time,as an important branch in the information security field,information hiding is also one of the research hot spots.Due to the wide applicability of the lightweight and highly condensed text,text-based information hiding technology is a difficult point which the scholars love to challenge.Tang Poetry is an cultural treasures of the Chinese nation and a special genre with great language change,high condensed and rich emotion.Text information hiding based on generating the Tang Poetry intelligently,this paper uses the theory and technology of deep learning to generate the Tang Poetry,and designs the information hiding algorithm under the combination of the image vocabulary of the Tang Poetry.The specific works are as follows:The researcher has collected a total number of 303440 ancient poetry works from the Tang Dynasty to the Qing Dynasty,which was used to form the initial corpus of Tang Poetry.On the basis of Shi Xue Han Ying,a modified dictionary of 1024 image-related two-character words was obtained by using the statistical the semantic similarity Sim method and semantic relevance correlation method.In the process of phonetic annotation,level and oblique tones marking of the modified dictionaries,the dictionaries of level and oblique tones are obtained.Then,through a further screening,a secret dictionary with 512 kinds of images vocabulary was obtained.On the theoretical basis of the circulatory neural network language model,in the operating system of Cent OS,using Tensorflow1.0 and Python2.7 opened by Google to build a system that can generate Tang Poetry intelligently,which was made up of a 5-layer and each layer containing 128 neurons of the recurrent neural network language model.This system mainly consists of Word2 Vec module,Train module and generated secret Tang Poetry module.After training,the system can generate better Tang Poetry.To design the text information hiding algorithm which can generate Tang Poetryintelligently.On the basis of the network model which was trained after the Tang Poetry corpus,according to the secret bit string information,transforming 9 bits into decimal number and then indexing to get the image in the secret dictionary.In order to complete information hiding,image is used as the keywords to generate every poem.In the generation process of poetry sentences,choosing the words from the secret dictionary that meet the requirements of level and oblique tones,and then replacing the words in the corresponding position,thus finally generating the secret Tang Poetry.Each poem can hide 9 bits of information,and The hidden capacity of the five-character Tang Poetry is 11.25%,and that of the seven-character Tang Poetry is 8.04%.This algorithm has high hidden quantity and concealment.
Keywords/Search Tags:text information hiding, deep learning, recurrent neural network, natural language processing, Tang Poetry
PDF Full Text Request
Related items