Unlock instant, AI-driven research and patent intelligence for your innovation.

Text detection method based on contour modeling

A text detection and outline technology, applied in character and pattern recognition, biological neural network models, instruments, etc., can solve problems such as difficulty in distinguishing adjacent text, poor effect of long text, and limited receptive field

Active Publication Date: 2021-06-08
HEFEI INSTITUTES OF PHYSICAL SCIENCE - CHINESE ACAD OF SCI
View PDF7 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In the field of text detection, the most commonly used algorithms are based on deep learning methods: there are two main methods, one is based on segmentation, and the other is based on detection; the former can detect text of any shape, but adjacent The text is difficult to distinguish, and the post-processing is complicated; the latter has a simple process, but cannot detect arbitrary shapes, and is limited by the receptive field, and is not effective for long texts

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
  • Text detection method based on contour modeling
  • Text detection method based on contour modeling

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0059] In this example, if figure 1 Shown, a kind of text detection method based on outline modeling is characterized in that, comprises the following steps:

[0060] Step 1: Collect pictures containing text and manually label them to create a picture dataset for text detection;

[0061] Step 2: Build a text detection model based on contour modeling, including: feature extraction module Backbone, feature fusion module FFM, text contour prediction module CM, and text coordinate offset prediction module OM;

[0062] The feature extraction module Backbone includes: 4 groups of BasicBlock layers, a convolution layer and 5 maximum pooling layers; each group of BasicBlock layers is composed of 3M convolutions;

[0063] The feature fusion module FFM includes 8 convolutional layers and 4 upsampling layers;

[0064] The text contour detection module CM includes N+1 convolutional layers;

[0065] The text coordinate offset prediction module OM includes N+1 convolutional layers;

[0...

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 discloses a text detection method based on contour modeling, and the method comprises the following steps: firstly collecting a picture containing characters, carrying out the manual marking, and building a picture data set for text detection; then, constructing a text detection model based on contour modeling; then running a text detection model based on contour modeling to obtain a text contour prediction result and a text coordinate offset prediction result; and finally performing post-processing to obtain a circumscribed polygon of each text. The invention can detect the text in any shape, and has the advantages of high robustness and high speed.

Description

technical field [0001] The present invention relates to relevant issues related to the field of text detection, in particular to a text detection method based on contour modeling. Background technique [0002] In the field of text detection, the most commonly used algorithms are based on deep learning methods: there are two main methods, one is based on segmentation, and the other is based on detection; the former can detect text of any shape, but adjacent The text is difficult to distinguish, and the post-processing is complicated; the latter has a simple process, but cannot detect arbitrary shapes, and is limited by the receptive field, which is not effective for long texts. Contents of the invention [0003] The present invention overcomes the shortcomings of the prior art, and provides a text detection method based on contour modeling, in order to determine text instances through contour prediction, thereby improving the accuracy and robustness of text detection. [0...

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): G06K9/32G06K9/46G06K9/62G06N3/04
CPCG06V10/255G06V20/62G06V10/454G06V10/44G06N3/045G06F18/253
Inventor 汪增福吴仕莲
Owner HEFEI INSTITUTES OF PHYSICAL SCIENCE - CHINESE ACAD OF SCI
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com