Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A Method for Localization, Segmentation and Recognition of Chinese in Natural Scene Images

A natural scene image and recognition method technology, applied in the field of Chinese positioning, can solve the problem that the model is difficult to expand

Active Publication Date: 2020-07-07
厦门商集网络科技有限责任公司
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method often requires a large amount of manually labeled data for training, and it is difficult to directly extend the trained model to more other application scenarios.

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
  • A Method for Localization, Segmentation and Recognition of Chinese in Natural Scene Images
  • A Method for Localization, Segmentation and Recognition of Chinese in Natural Scene Images

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] Such as figure 1 As shown, this embodiment provides a Chinese positioning, segmentation and recognition method in natural scene images, and the process can be divided into the following steps:

[0024] 1) Perform preliminary text positioning on the original image through the FASText model, and extract candidate text areas;

[0025] 2) By pre-segmenting the candidate text area;

[0026] 3) Recognize the word part of the pre-segmented text area, and perform further word segmentation and recognition on the field part.

[0027] Such as figure 2 Figure (a) is the original picture; step 1 uses the getCharSegmentation function of FASText to extract the candidate image area, and the extracted image is shown in figure (b); the pre-segmentation operation of step 2 is to first determine the Unicom extracted in step 1 area, after removing some smaller Unicom areas (noise), consider the area that conforms to the aspect ratio of Chinese characters (close to 1:1) as a single chara...

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 a Chinese positioning, segmentation and recognition method in natural scene images. The present invention performs preliminary text positioning on the original picture through the FASText model, extracts candidate text areas, pre-segments the candidate text areas, and then recognizes the single-character part of the pre-segmented text area, and further performs field part Word segmentation and recognition. The present invention utilizes the accurate extraction of character stroke features, and the powerful character recognition ability of the deep residual neural network, combined with the path tree method, to achieve the purpose of Chinese positioning and recognition simply and effectively, and can be applied to various natural scenes without supervision and training .

Description

technical field [0001] The invention belongs to the technical field of image processing, and in particular relates to a Chinese positioning, segmentation and recognition method in natural scene images. Background technique [0002] Text recognition in natural scenes is a very important visual detection target. The text in the image contains a lot of useful information, which is crucial to the understanding and acquisition of visual content. At present, there are many related text recognition applications, including road signs, license plates, bills and so on. [0003] Generally speaking, traditional OCR technology is affected by the complex background of natural scenes, and it is difficult to complete related tasks correctly. Overall, such tasks can be divided into two stages, text location and recognition. Text localization is the precise location of text in an image, mainly based on extracting relevant text features, such as MSERs, to distinguish fields from backgrounds....

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 Patents(China)
IPC IPC(8): G06K9/20G06K9/32G06K9/34
CPCG06V10/22G06V10/245G06V10/267
Inventor 陈凯韦建何建华周异黄征杜保发周文贵查宏远
Owner 厦门商集网络科技有限责任公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com