Mathematical formula identification method, device and equipment

Abstract

The invention belongs to the field of optical character recognition, and particularly relates to a mathematical formula recognition method, device and equipment. The method comprises the following steps: acquiring a mathematical formula picture to be identified, and preprocessing the mathematical formula picture; distinguishing a plurality of rows of mathematical formula pictures and a single rowof mathematical formula pictures from the mathematical formula pictures to be identified by adopting a LeNet classifier; cutting the multiple rows of mathematical formula pictures into multiple single-row mathematical formula pictures in a projection mode; segmenting all the single-row mathematical formula pictures into single mathematical character pictures; identifying the type of each mathematical character picture by adopting a neural network; utilizing an improved baseline identification method to identify the relative position between the single mathematical characters; and forming a semantic tree of the mathematical formula picture to be identified, analyzing the semantic tree into a latex language, and outputting the latex language. The method has the advantages of being high in recognition rate of recognition formula characters and complete in structure recognition.

Description

technical field [0001] The invention belongs to the field of optical character recognition, and relates to the recognition of mathematical formulas by combining deep learning and rule learning. The invention aims to provide a method for formula recognition with high character recognition accuracy and formula recognition accuracy; specifically, a A recognition method, device and equipment for a mathematical formula. Background technique [0002] After decades of development, Optical Character Recognition (OCR) technology has produced a series of relatively mature products at home and abroad, such as Tsinghua Unisplendour OCR, Hanwang OCR, Saiku OCR and foreign Taibi OCR. These OCR product technologies have reached a very high level of recognition of Chinese, English and Arabic numerals contained in printed books or documents, but they cannot recognize mathematical formulas well. The main reasons include: compared with conventional characters, mathematical formulas have a com...

AI Technical Summary

Problems solved by technology

This will cause two problems: one is to increase the overhead of storage and transmission; the other is to fail to realize the reuse of formulas

[0008] But in general, the above methods have their limitations. For the identification method of constructing the symbol center baseline, the identification method with high utilization rate (Zhang X, Gao L, Yuan K, et al.A Symbol Dominance Based Formulae Recognition Approach for PDF Documents[C] / / 2017 14th IAPRInternational Conference on Document Analysis and Recognition(ICDAR).IEEEComputer Society,2017.) as an example, the recognition method using the baseline method must first determine the start symbol, which is generally considered is the leftmost character in the formula, but the leftmost character of the entire mathematical formula cannot determine the center reference line of the entire formula, which may lead to a wrong determination of the center reference line, and the entire formula structure will be misrecognized

Images