General method for extracting document structural information

Abstract

The invention provides a general method for extracting document structural information and belongs to the field of document integration projects. According to domain knowledge of a specific domain, an isomorphic information format of a document is defined, the isomorphic information format at least comprises a text node for defining a text information format, a structure node for defining a document structural information format, a picture node for defining a picture information format and a table node for defining a table information format, an extraction and conversion method of an original document to the isomorphic information format, an unified operation interface used by an upper layer and an isomorphic information description format are built, and the isomorphic information format is converted into the isomorphic information description format for displaying. The method can be used for extracting the important document structural information, meanwhile keeps pictures and table information in the document, and is simple and easy to use and strong in universality. By means of the method, a document isomorphic information model in the specific domain can be built, isomorphic interoperability of document information is achieved, and document integrated management is convenient.

Description

technical field [0001] The invention belongs to the field of document integration engineering, and relates to a general document information format definition, a method for realizing the conversion and operation process between a group of document information formats, and a general definition of document isomorphic information description; in particular, it relates to a general Document structure information extraction method. Background technique [0002] With the development of related technologies, the status of document resources in engineering practice has become more and more prominent. The concept of document engineering put forward in the new century puts the status of document resources at the center of engineering practice. Document resources are a kind of knowledge accumulation and the crystallization of experience in engineering practice. Making full use of existing document resources can reduce mistakes made in current engineering practice, provide reference fo...

AI Technical Summary

Problems solved by technology

However, the research objects of these research works are only locked on the text content and the semantics it represents, ignoring the original document structure information and chart information. In addition, because its architecture is too general, there is no specific field-oriented The isomorphic technology and method of extracting operable documents makes it doomed to fail to meet the actual needs of engineering

[0006] To sum up, there are many limitations in the existing document integration research field: 1) In document information extraction, only focus on the extraction of text information, ignoring document structure information, so that the extracted plain text is convenient for text retrieval and classification, etc. At the same time of application, due to the lack of important structural information, it cannot meet the needs of specific engineering fields; 2) In the document information extraction, the important picture and chart information in the document is ignored, which is convenient for the definition of the isomorphic format of the general document. There is information on document resources, but the original document resources are not fully utilized; 3) When defining an open document isomorphic structure, the concept of an open document hierarchical model is proposed, and key technologies for obtaining text information in multiple formats are introduced and methods, but did not give a domain-specific isomorphic document format and its definition method, and did not propose a practical process and method for establishing an open document isomorphism for a specific domain; 4) open documents Isomorphism mainly studies text information extraction processing and semantic understanding. There is no general document information description method, which cannot be understood and operated by people, and cannot meet the actual needs of engineering; It is easy to popularize in engineering practice; 6) The extraction method has poor versatility and cannot guarantee portability

Images