A method for recognizing a text in a clothing tag image

A Chinese text and image technology, applied in character and pattern recognition, instruments, computer components, etc., can solve problems that have not been reported, and achieve the effects of reducing error rate, time and cost compression, and high recognition accuracy

Pending Publication Date: 2019-06-21
上海宝尊电子商务有限公司
View PDF5 Cites 31 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] But about the method for text recognition in a kind of clothing tag image of the present invention, there is no report yet

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 recognizing a text in a clothing tag image
  • A method for recognizing a text in a clothing tag image
  • A method for recognizing a text in a clothing tag image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0031] Please refer to the attached figure 1 , with figure 1 It is a flow chart of a method for text recognition in a clothing tag image in this embodiment.

[0032] The method for text recognition in the clothing tag image is a tag recognition scheme based on deep learning technology, and specifically needs to carry out four processes;

[0033] First: Correct the angle of the text in the image. Since the position of the hang tag in the picture uploaded by the user will inevitably be tilted, it is necessary to detect the tilt angle of the hang tag area and correct it for rotation;

[0034] Second: Further detect the text area in the picture. The target detection technology is used to detect the area where there is text, and the input image is cropped. This preprocessing can reduce the time-consuming of subsequent text line synthesis and improve the detection accuracy. Then locate the text lines in the image for subsequent recognition of line-by-line text;

[0035] Third: ...

Embodiment 2

[0039] Please refer to the attached figure 2 , with figure 2 It is a schematic diagram of a method for text recognition in a clothing tag image of this embodiment.

[0040] The method for text recognition in the clothing tag image comprises angle correction (1), text detection (2), text recognition (3), text correction (4); , get rid of outliers, calculate the average value as the rotation angle of the picture, and carry out rotation correction to it; described text detection (2) calculate the edge position of the character area by the mode of tag area detection, carry out picture clipping and then carry out text line detection, With simple text line construction algorithm, the character frame that target detection classification obtains is merged into text line, obtains the coordinate of text line; Convolutional Recurrent Network (CRNN) combined with RNN) realizes text recognition. Among them, the convolutional layer is used for feature extraction, the recurrent layer is...

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 relates to a method for recognizing a text in a clothing tag image. The method comprises the steps of angle correction, text detection, text recognition and text correction. The angle correction includes carrying out rotation correction by calculating the rotation angle of the picture; text detection includes: calculating the edge position of the character area for picture clipping,and then using a simple text line construction algorithm to obtain coordinates of a text line; Text recognition is achieved through a convolutional cycle network (CRNN), and character classification is achieved through images under different fonts, character numbers and backgrounds which are synthesized through codes and part of real scene pictures. And text correction includes carrying out word bank matching and replacement by vocabularies with the highest similarity, and correcting coherent entries. The method has the advantages that a traditional manual information input mode can be replaced, the boring and cumbersome manual input process is reduced, meanwhile, extreme compression is achieved in time and cost, the recognition result achieves high recognition precision, and the word bankmatching correction reduces the error rate.

Description

technical field [0001] The invention relates to the technical fields of target detection and text recognition, in particular to a method for text recognition in clothing tag images. Background technique [0002] Optical Character Recognition (English: Optical Character Recognition, OCR) traditionally refers to the process of analyzing and recognizing scanned document images to obtain text and layout information. This technology has also been further applied to text recognition in scenes, that is, to recognize text information in natural scene images. Due to the extremely diverse backgrounds in natural scene images, the text allows multiple languages ​​to be mixed, and the characters can have different sizes, fonts, colors, etc., and the text area in the image may be deformed, incomplete, blurred, etc., the difficulty of character recognition in the scene is much greater than that of traditional methods. Text recognition in document images. At present, OCR technology is wid...

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): G06K9/20G06K9/32
Inventor 胡玉琛吴磊彬林博
Owner 上海宝尊电子商务有限公司
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