Complex PDF structure analysis method and device based on neural network

Abstract

Embodiments of the invention provide a complex PDF structure analysis method and device based on a neural network. The method comprises the steps of obtaining feature information of a PDF document; carrying out coarse particle division on the feature information of the PDF document according to a maximum entropy model to obtain a layered paragraph of the PDF document; converting layered paragraphsof the PDF document according to a two-layer bidirectional language model trained in a large-scale corpus set to obtain paragraph word vectors, and compressing the paragraph word vectors to obtain paragraph semantic vectors; and inputting the paragraph semantic vector into a multi-layer bidirectional long-short-term memory network to obtain a hierarchical sequence of all paragraphs of the PDF document. The technical problem that the generalization ability is poor due to the fact that the PDF document structure is not single is solved, and the technical effects that the limitation of manual design rule logic is avoided, the complex PDF document structure can be analyzed at a high level, and the generalization ability is high are achieved.

Description

technical field [0001] The embodiments of this specification relate to the field of computer technology, and in particular to a neural network-based complex PDF structure analysis method and device. Background technique [0002] The PDF document parsing method is mainly used to establish the PDF document structure system, and can also prepare for further extraction of document entity information. PDF is a common file format. One is a document with a clear directory structure and text distinction. This form often requires manual entry and typesetting. The other is the vast majority of PDF documents, which are scanned pictures of physical originals one by one. There is neither a directory structure nor a clear text distinction for page saving, which is not conducive to reading or further information extraction. The current mainstream PDF document structure extraction method is based on the document content and design rules to extract several titles that may be used as directo...

AI Technical Summary

Problems solved by technology

[0005] The embodiment of this specification provides a complex PDF structure analysis method and device based on a neural network, which solves the problem that in the prior art, because the PDF document structure is not single, it is difficult to design analysis rules that can adapt to various structures. The technical problem of poor generalization ability has achieved the limitation of avoiding the logic of manual design rules, and has a higher level in parsing different PDF documents and complex PDF structures, and the technical effect of strong generalization

Images