Research On Zero-Watermarking Algorithm For Argumentative Chinese Text

With the deep integration of the Internet and various fields of society,a large number of digital works with text as the main carrier are widely disseminated,promoted and exchanged through the Internet.However,due to the openness and sharing of network communication and the easy copying and tampering of text information itself,the copyright protection of text digital works is becoming increasingly urgent.One of the effective ways to solve this problem is to use text watermarking technology.Aiming at the problems of weak text representation,ignoring stylistic diversity and low accuracy of watermarking judgment in existing text watermarking schemes,this paper designs a zero-watermarking scheme for Chinese text based on argumentation according to the characteristics of argumentation,and carries out in-depth discussion and Research on this scheme.The main work includes the following aspects:(1)Based on the analysis and summary of existing text watermarking schemes,a zero-watermarking scheme for Chinese text is designed according to the characteristics of argumentation.The scheme mainly includes text preprocessing,text zero-watermarking construction and text zero-watermarking detection.(2)In the part of text preprocessing,the preprocessing model and specific workflow are designed,and the algorithms of key modules in the model,such as document structure recognition and word segmentation,word frequency statistics,subject word synonym extraction and word weight calculation,are designed and implemented.(3)In the part of zero-watermarking construction of text,a zero-watermarking construction model and its specific workflow are designed based on the thematic nature of general style and the characteristics of argumentation putting facts and reasoning.The algorithms of key modules in the model are designed and implemented,such as extraction of paragraph center sentences,extraction of theoretical facts,extraction of logical relations of articles,feature fusion and encryption.(4)In the zero watermark detection part of the text,the method of watermark information similarity calculation is used to determine the copyright.The model and specific workflow of zero watermark detection and determination are designed,and the algorithms of key modules in the model are designed and implemented,such as archive zero-watermarking parsing,paragraph center sentence similarity calculation,article logical relationship similarity calculation and copyright determination.(5)The performance and anti-attack of the proposed scheme are analyzed and compared.The results show that the proposed scheme has strong pertinence and applicability for copyright protection of argumentative text.