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

Natural scene text detection method and system based on non-local network

A natural scene and text detection technology, applied in the field of image processing, can solve the problem of misjudging the background area as a text object, etc., and achieve the effect of reducing the resolution of the feature map, strengthening the connection, and reducing the number of channels

Active Publication Date: 2019-12-17
HUAZHONG UNIV OF SCI & TECH
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a method and system for text detection in natural scenes based on non-local networks, thereby solving the technical problem of misjudging the background area as a text object in the prior art

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
  • Natural scene text detection method and system based on non-local network
  • Natural scene text detection method and system based on non-local network
  • Natural scene text detection method and system based on non-local network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] In order to make the object, technical solution and advantages of the present invention more clear, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not constitute a conflict with each other.

[0044] The invention provides a method for detecting text in natural scenes based on a non-local network. The purpose is to reduce the generation of false positive samples by adding the global information of the picture to the region of interest in the feature extraction stage by constructing a non-local network.

[0045] Such as figure 1 As shown, a non-local network-based natural scene tex...

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 natural scene text detection method and system based on a non-local network, and belongs to the technical field of image processing, and the method comprises the steps: extracting the multi-scale features of a picture, reconstructing each scale feature through the non-local network, and obtaining the reconstructed features; generating a candidate region of the text by utilizing the reconstructed features; and extracting a region of interest according to the candidate region, performing text and non-text classification and position and size adjustment on the region ofinterest through a classification and regression network, and finally obtaining a text target in the image. According to the method, the non-local network is used for reconstructing each scale feature, so that the occurrence of a pseudo-positive sample in a text detection result can be effectively inhibited, and the technical problem that a background region is misjudged as a text target in the prior art is solved.

Description

technical field [0001] The invention belongs to the technical field of image processing, and more specifically relates to a method and system for detecting text in natural scenes based on non-local networks. Background technique [0002] As a form of expression of text, text in natural scenes has rich and precise high-level semantic information, which plays a very important role in understanding the surrounding scenes. Because of this, the automatic acquisition of text information in natural scenes has broad application prospects in areas such as autonomous driving, instant translation, robot navigation, and industrial automation. How to accurately and efficiently extract text information in natural scenes has become one of the current hot issues in computer vision. This task includes two sub-tasks, text detection and text recognition. The former is the premise of realizing the latter. [0003] In the existing two-stage natural scene text detection algorithm, after obtainin...

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): G06K9/00G06K9/32G06K9/34G06K9/46G06K9/62
CPCG06V20/10G06V20/63G06V30/158G06V10/464G06V30/10G06F18/2415G06F18/241
Inventor 姜军解猛周作禹胡忠冰曾雄梅胡若澜
Owner HUAZHONG UNIV OF SCI & TECH