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

Text string recognition method and device and electronic equipment

A text recognition and text technology, which is applied in the field of recognition processing, can solve the problems of lower recognition accuracy of text strings, achieve the effect of reducing the amount of calculation and realizing simplicity

Pending Publication Date: 2020-12-25
SHANGHAI GOLDWAY INTELLIGENT TRANSPORTATION SYST CO LTD
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when the text strings are dense in the cutout image, other text string fragments will appear in the cut out image area, and the text string recognition accuracy will be reduced.

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 string recognition method and device and electronic equipment
  • Text string recognition method and device and electronic equipment
  • Text string recognition method and device and electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0067] In Embodiment 1, the trained text recognition model may include a first spatial attention module. Here, the first spatial attention module is just named for convenience of description, and is not used for limitation.

[0068] In Embodiment 1, the first spatial attention module is used to obtain channel feature maps of S*S channels based on the trained text string position supervision information and based on the input image feature maps. Among them, the S*S channels are the channels configured by the first spatial attention module.

[0069] Based on this, determining the channel feature maps of S*S channels according to the image feature maps of the target image in the above step 101 may include: inputting the image feature maps of the target images into the first spatial attention module in the trained text recognition model To get the channel feature map of S*S channels.

[0070] The following describes how the first spatial attention module obtains channel feature ...

Embodiment 2

[0081] The above describes how the first spatial attention module in Embodiment 1 learns the channel feature maps of S*S channels based on the input image feature maps. Embodiment 2 is described below:

[0082] Example 2:

[0083] In Embodiment 2, the trained text recognition model may include a second spatial attention module. Here, the second spatial attention module is just named for convenience of description, and is not used for limitation.

[0084] In Embodiment 2, the second spatial attention module is used to determine channel feature maps of S*S channels based on the trained word position supervision information and based on the input image feature maps. In Embodiment 2, the channel feature map of each channel includes L word position segmentation maps, and each word position segmentation map is used to characterize the word position. In the second embodiment, the word positions corresponding to the L word position segmentation maps included in each channel feature...

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 text string recognition method and device and electronic equipment. A target image is divided (also called grid drawing), and channel feature maps (used for representing positions of text strings associated with image blocks corresponding to the channels) of S * S channels are determined according to image feature maps of the target image; Each text string in the target image is mapped to different channels to learn the positions of the text strings in a grid drawing mode, and finally text recognition is realized on the premise of not using matting operation, so that compared with the matting operation, the calculation amount can be greatly reduced.

Description

technical field [0001] The present application relates to recognition processing technology, in particular to a text string recognition method, device and electronic equipment. Background technique [0002] Text string recognition is an important application of pattern recognition. The text string here, from one aspect, it can be a text, a sentence (Sentence), a paragraph (Paragraph) or a chapter (Discourse) etc.; From another aspect, it can also be a character, A symbol, a string of the same or different characters and / or symbols, etc. The text strings in this application may not be limited to the text interpretation in the current literature field, and this application does not limit the specific form of the text. [0003] Text string recognition can be applied to many business fields, such as: reading, translation, literature retrieval, statistical summary of commodity invoices, commodity code recognition, commodity warehouse management, certificate recognition, license...

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
IPC IPC(8): G06K9/34G06K9/46G06K9/20G06K9/62
CPCG06V10/22G06V30/153G06V10/40G06V30/10G06F18/253
Inventor 乔梁
Owner SHANGHAI GOLDWAY INTELLIGENT TRANSPORTATION SYST CO LTD
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