Text image correction method and device, computer device and storage medium

Abstract

The invention relates to a text image correction method and device, a computer device and a storage medium. The method comprises the steps of obtaining a text image needing to be corrected; inputtinga text image to be corrected into the edge detection model for deep learning to obtain a binary edge image; wherein the edge detection model is obtained by training a deep learning network through a plurality of text images with edge information; performing text fine positioning on the binary edge image to obtain a text bounding box and vertex coordinates of the text bounding box; calculating a perspective transformation matrix according to the vertex coordinates of the text bounding box; performing perspective transformation on the image in the text bounding box and the vertex coordinates ofthe text bounding box according to the perspective transformation matrix to obtain a corrected text image; and sending the corrected text image to the terminal for the terminal to consult and call. According to the method, the document edge extraction precision under different scenes is realized, the interference of noise edges of non-document regions is reduced, and the document coordinate position is refined.

Description

technical field [0001] The present invention relates to an image processing method, and more specifically refers to a text image correction method, device, computer equipment and storage medium. Background technique [0002] Document digitization plays an important role in the intelligent analysis and processing of document content and the preservation of precious literature materials, and is the basic means of digital government construction. Traditional document digitization operations mainly use special scanners to scan, which usually have disadvantages such as not easy to carry and high cost. In recent years, with the continuous popularization of smart phones, more and more people have begun to use mobile phone cameras to scan common documents such as books and bills in life, and then automatically complete the digital scanning of documents through optical character recognition technology and natural language processing technology and analysis. However, when using a po...

AI Technical Summary

Problems solved by technology

The traditional technical means mainly use the traditional Canny edge detection operator to extract the edge information of the image, and design a series of rule judgment algorithms to obtain the bounding box that conforms to the characteristics of the document image. There are various problems in the edge information obtained by the Canny algorithm, including Edge segments of different lengths and complex background edge noise, etc., the existing method is mainly to continuously adjust the detection threshold of the canny operator to adapt to the document image in different scenarios, but the parameter adjustment method based on manual experience will be less robust. Weak problem, different scenarios may need to set different parameters, with the great progress of deep learning in target detection, segmentation and other visual fields, the target detection algorithm based on YOLO is widely used in actual scenes, however, the output of YOLO algorithm It is usually a non-rotated rectangular box representing the target location area, which cannot be positioned for document areas with angles

Images