Method and system for detecting form frame lines in form documents

Abstract

The invention discloses a method and a system for detecting a form frame line in a form document. The method comprises the following steps: S1, acquiring an image of a shot paper form document; S2, calculating an inclination angle of the image and performing angle correction on the image to enable a table in the image to be in a horizontal and vertical state; S3, detecting the corrected image by using a table line segment intersection detection model based on deep learning, and recording the position and type of a table line segment intersection in the image; S4, processing the intersection points by using the positions and types of the intersection points of the table line segments, removing the mistakenly identified intersection points, and supplementing the missing and unidentified intersection points; and S5, accurately positioning the final effective intersection point position, and connecting the adjacent intersection points in the horizontal direction and the vertical directionaccording to the intersection point position and type so as to recover the table shape. According to the invention, automatic intelligent detection of the form frame line can be realized, so that theline segment intersection point in the form can be quickly and accurately identified, the form shape is recovered, and a basis is provided for next text analysis of documents.

Description

Technical field [0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and system for detecting form borders in form documents. Background technique [0002] At present, the analysis and inspection of paper forms are basically performed manually, and the contents of the forms are confirmed by manual reading. Manual review has the characteristics of low efficiency, error-proneness, high labor cost, etc., and cannot realize the automation of document review. Manual review is not conducive to the business development and management of various institutions. [0003] Using a computer to quickly and accurately identify paper forms and documents has become an urgent problem to be solved. Among them, the detection of form borders has become the first step in the intelligent identification of form documents, but there is no effective solution in the prior art to realize the form frame. Automatic intelligent detection of lines...

AI Technical Summary

Problems solved by technology

Manual review has the characteristics of low efficiency, error-prone, high labor costs, etc., and cannot realize the automation of document approval

Manual review is not conducive to the business development and management of various institutions

[0003] Using computers to quickly and accurately identify paper form documents has become an urgent problem to be solved. Among them, the detection of form frame lines has become the first step in the intelligent identification of form documents, but there is no effective solution in the prior art to realize the form frame. Automatic intelligent detection of lines

Images