An OCR method based on depth learning

Abstract

The invention discloses an OCR method based on depth learning, comprising the following steps of obtaining an image to be recognized; scaling the image to be recognized, and then preprocessing the scaled image, wherein the preprocessing comprises sharpening, graying, binarization, correcting tilt, reducing noise, and removing any one or more items in the official seal; putting the preprocessed image into the depth learning neural network to obtain a plurality of candidate text regions of the scaled image; putting a plurality of candidate text regions of the scaled image into a depth-based learning neural network, and converting the contents of the candidate text regions into computer-readable text; putting the computer-readable text into the deep learning language model to correct the textrecognition results. The method of the invention can realize the intelligent extraction of the text content of the image, provides the service of fast searching and fast obtaining information for thepersonnel of various trades, improves the working efficiency, has high identification field accuracy rate and improves the processing speed.

Description

technical field [0001] The invention relates to the technical field of image recognition in computer software, in particular to an OCR method based on deep learning. Background technique [0002] The text area detection and positioning recognition technology based on optical character recognition (hereinafter referred to as OCR) refers to the use of computers and other equipment to automatically extract and recognize valid information in paper materials and perform corresponding processing. One of the key technologies of paper-based computer automatic processing. The traditional image text recognition is OCR, which is based on scanning the paper documents to be recognized into electronic images. However, considering the quality of the scanning effect, the quality of the paper document itself (such as printing quality, font clarity, font standardization, etc.), and the content layout (the arrangement of the text, compared with ordinary text and table text and bills), The ac...

AI Technical Summary

Problems solved by technology

However, considering the quality of the scanning effect, the quality of the paper document itself (such as printing quality, font clarity, font standardization, etc.), and the content layout (the arrangement of the text, compared with ordinary text and table text and bills), The actual effect of OCR is not always satisfactory

The requirements for recognition accuracy of different paper documents are different. For example, the recognition of bills has very high requirements for accuracy, because if a digital recognition error may lead to fatal consequences, traditional OCR recognition cannot meet the requirements. Such high-precision identification requirements

Images