In order to put an end to academic misconduct, the requirements for duplicate checking of papers are becoming more and more strict. As we all know, a paper generally consists of title, table of contents, abstract, keywords, introduction, body, conclusion, references and other parts. Moreover, the contents of paper duplicate checking mainly include word count, paper format and paragraph, chapter conversion, text, abstract and conclusion, while non-text parts such as formulas, charts and pictures in the paper do not participate in duplicate checking. Simply put, the text in the paper is the main object of duplicate checking. Usually, the system will check the content of the paper and then check it chapter by chapter. The comparison scope of the search includes all kinds of published journals, conference papers, patent contents, graduate papers over the years, Internet resources, etc.