Multidirectional text comparison method

Abstract

The invention provides a multidirectional text comparison method. The method comprises the following steps: exporting a package design drawing into a PDF format, and analyzing text content and corresponding position information; splitting a text analyzed by the PDF according to intervals, and judging whether split texts are consistent or not; calculating 2-gram word frequency through a large number of Chinese corpora, calculating a probability of a positive sequence of the text and a probability of a negative sequence of a text, and judging whether the texts are in the negative sequence or not according to the probabilities; performing block processing on all texts according to position coordinates, judging directions of text blocks according to the positive and negative sequences of the texts, and then sorting and combining the texts in the text blocks; and comparing the PDF text content with standard text content for examination, matching similar lines, and labeling differences of the similar lines. According to a processing result provided by the invention, structural differences between an analysis text and an actual text can be greatly reduced, the detection precision of the analysis text and the actual text is improved, and the workload of manual intervention is reduced, so that the packaging design cost and precision are reduced.

Description

technical field [0001] The present invention relates to the field of image processing, and more specifically, the present invention relates to a text direction correction and text comparison method based on machine learning. Background technique [0002] In the field of packaging design, in order to ensure the accuracy of the text information in the packaging design drawing, it is usually necessary to review the text content in the design drawing before finalizing it, especially important information such as various license numbers. irreversible loss. [0003] Conventional text error comparison is mostly done manually, but the manual method not only has a large labor and time cost, but also is prone to problems such as missed detection and false detection. Since the printed design drawings can be analyzed to obtain the text content, some manufacturers will use machine learning methods for comparison, which can achieve the desired effect in other printing design fields. How...

AI Technical Summary

Problems solved by technology

However, the disassembly diagram of the package will be designed in the packaging design drawing, which will have many text textures rotated at different angles. When parsing the design drawing file, it will cause problems such as reverse order, branching, and overlapping of the text. , direct comparison will cause a large amount of texts to fail to be compared successfully, affecting the comparison result

Images