Image correction and text and position identification method and system

A technology of image correction and recognition methods, applied in the field of image vision

Active Publication Date: 2019-07-09
BEIJING UNION UNIVERSITY
View PDF9 Cites 23 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In order to solve the above-mentioned technical problems, the present invention proposes a method and system for image correction and text and position recognition, based on a neural network image correction and text and position recognition model, which mainly solves the problem of text and its position in ID cards, business cards, form pictures, etc. Identify problems to meet the application needs of various industries and bring better experience to users

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image correction and text and position identification method and system
  • Image correction and text and position identification method and system
  • Image correction and text and position identification method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0084] Such as figure 1 , 2 As shown, step 100 is executed, and the input module 200 inputs the picture to be detected.

[0085] Step 110 is executed, the detection module 210 detects the text angle of the picture to be detected, fits a straight line, and counts the slope of the straight line, and uses the mode direction θ as the correction direction of the picture. Using the duality relationship between the point and the line, the discrete points in the image space are transformed into curves in the Hough space, and the intersection points of the curves are used as the parameters of the straight line equation, and the parameters are counted. The formula for the conversion is as follows:

[0086] ρ=x 1 cosθ+y 1 sinθ, where ρ represents the representation of pixels in Hough space, x 1 Indicates the abscissa of the pixel in the image space, y 1 Indicates the vertical coordinate of the pixel in the image space. Count the intersection points of the curves converted to Hou...

Embodiment 2

[0093] An image rectification and text and position recognition model method, comprising the following steps:

[0094] The first step, for the input picture (such as image 3 As shown), detect the text angle of the picture, fit the straight line, and count the slope of the straight line, and use the mode direction as the correction direction of the picture. Using the dual relationship between points and lines, the discrete points in the image space are converted into curves in Hough space, and the intersection points of the curves are used as parameters of the straight line equation. The conversion equation is as follows:

[0095]

[0096] Statistically convert the intersection point of the curve into Hough space. If it exceeds the threshold, it is considered as the text direction, and the parameters (ρ, θ) are recorded, and the mode of the parameter is further counted, and θ is used as the rotation angle.

[0097] The second step is to use the affine transformation matrix...

Embodiment 3

[0116] This patent proposes an image-based text information and its position detection and recognition system OCR (optical character recognition) to meet the application needs of various industries and bring better experience to users. OCR (optical character recognition) is one of the applications of image-based sequence recognition. Image-based sequence recognition has been a long-term research topic in the field of computer vision. Characters, and then the process of translating the shape into computer text by character recognition; that is, the process of scanning the text data, and then analyzing and processing the image file to obtain the text and layout information. In order to better apply OCR technology to different scene recognition, firstly, the image to be detected is rotated to improve the accuracy of target area detection in the neural network, thereby improving the accuracy of text recognition and detection. Pure text recognition technology cannot satisfy all OCR...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention provides an image correction and text and position recognition method and system, and the method comprises the steps: inputting a to-be-detected picture and also comprises the followingsteps of detecting character angles of the to-be-detected picture, fitting a straight line, counting the slope of the straight line, and taking mode direction Theta as the correction direction of thepicture; utilizing affine transformation matrix to rotate the position of a to-be-detected picture; pre-identifying information of the to-be-detected picture by using fast-rcnn positioning technology;inputting a pre-identified target area into convolutional deep neural network CLNN for accurate identification of characters and positions thereof; and outputting a recognition result. The inventionprovides the image correction and text and position identification method and system. According to the image correction and text and position identification model based on a neural network, the problems of identification of texts such as an identity card, a business card and a table picture and position identification of the texts are solved so as to meet the application requirements of various industries and bring better experience to a user.

Description

technical field [0001] The invention relates to the technical field of image vision, in particular to a method and system for image correction and text and position recognition. Background technique [0002] Text recognition and detection of image sequences are required in many industries and occasions, such as text detection of ID card information. Banks, railway stations, airports, hotels, etc. have specialized staff to carry out this work. The original intention of the research and development of the text and location detection and recognition system is based on the deep learning network, using deep features to represent ID card information, to achieve fast and accurate text recognition and detection. With the development of the mobile Internet, more and more application technologies involve the input authentication of credential information (namely real-name authentication), the speed of manually inputting information is relatively slow, and the user experience is poor. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/32G06K9/34G06K9/46G06N3/04
CPCG06V10/242G06V30/153G06V10/48G06N3/045
Inventor 何宁孙欣
Owner BEIJING UNION UNIVERSITY
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap