View in EDS PDF Full Text

DBNetText Detection Algorithm Based on Edge Detection.

Saved in:
Bibliographic Details
Title: DBNetText Detection Algorithm Based on Edge Detection.
Authors: FAN, Huiqiong1,2 fhq8109@163.com, WAN, Changxuan1,2 wanchangxuan@263.net
Source: Technical Gazette / Tehnički Vjesnik. 2025, Vol. 32 Issue 6, p2188-2198. 11p.
Subjects: Edge detection (Image processing), Text recognition, Electronic commerce, Image processing
Abstract: In the booming e-commerce industry, precise text detection in product images is crucial for seamless operations. However, existing text detection algorithms face challenges due to the complex nature of e-commerce images. These images often combine intricate text with complex graphics and diverse product elements, all set against highly variable backgrounds. Artistic fonts, with their unique and often ornate designs, are especially difficult to detect accurately, leading to subpar performance in extracting product-related information. This inefficiency limits the development of intelligent e-commerce applications, which motivates our research. To address these challenges, we propose EIEM-DBNet, an edge-detection-based text detection algorithm. Its key innovation is the integration of the Edge Information Extraction Module (EIEM), which uses operators like Laplace, Sobel, and Canny to extract edge details from low-level feature maps. By emphasizing local edge features, EIEM-DBNet better distinguishes text from the complex background compared to traditional methods that rely on global features. After edge detection, a channel-weighting mechanism incorporates the extracted edge information into the model, enhancing its text detection accuracy. In terms of performance, EIEM-DBNet outperforms traditional DBNet models. In ablation experiments, it shows a 1.1% increase in recall, a 1.3% rise in accuracy, and a 1.2% improvement in F1-score. Compared to other advanced models, EIEM-DBNet achieves the highest recall rate in terms of F1-score, indicating its superior ability to balance precision and recall, thereby providing more accurate text detection in complex e-commerce image scenarios. [ABSTRACT FROM AUTHOR]
Copyright of Technical Gazette / Tehnički Vjesnik is the property of Tehnicki Vjesnik and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Links:
  – Type: pdflink
Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 190580231
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: DBNetText Detection Algorithm Based on Edge Detection.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22FAN%2C+Huiqiong%22">FAN, Huiqiong</searchLink><relatesTo>1,2</relatesTo><i> fhq8109@163.com</i><br /><searchLink fieldCode="AR" term="%22WAN%2C+Changxuan%22">WAN, Changxuan</searchLink><relatesTo>1,2</relatesTo><i> wanchangxuan@263.net</i>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Technical+Gazette+%2F+Tehnički+Vjesnik%22">Technical Gazette / Tehnički Vjesnik</searchLink>. 2025, Vol. 32 Issue 6, p2188-2198. 11p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Edge+detection+%28Image+processing%29%22">Edge detection (Image processing)</searchLink><br /><searchLink fieldCode="DE" term="%22Text+recognition%22">Text recognition</searchLink><br /><searchLink fieldCode="DE" term="%22Electronic+commerce%22">Electronic commerce</searchLink><br /><searchLink fieldCode="DE" term="%22Image+processing%22">Image processing</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: In the booming e-commerce industry, precise text detection in product images is crucial for seamless operations. However, existing text detection algorithms face challenges due to the complex nature of e-commerce images. These images often combine intricate text with complex graphics and diverse product elements, all set against highly variable backgrounds. Artistic fonts, with their unique and often ornate designs, are especially difficult to detect accurately, leading to subpar performance in extracting product-related information. This inefficiency limits the development of intelligent e-commerce applications, which motivates our research. To address these challenges, we propose EIEM-DBNet, an edge-detection-based text detection algorithm. Its key innovation is the integration of the Edge Information Extraction Module (EIEM), which uses operators like Laplace, Sobel, and Canny to extract edge details from low-level feature maps. By emphasizing local edge features, EIEM-DBNet better distinguishes text from the complex background compared to traditional methods that rely on global features. After edge detection, a channel-weighting mechanism incorporates the extracted edge information into the model, enhancing its text detection accuracy. In terms of performance, EIEM-DBNet outperforms traditional DBNet models. In ablation experiments, it shows a 1.1% increase in recall, a 1.3% rise in accuracy, and a 1.2% improvement in F1-score. Compared to other advanced models, EIEM-DBNet achieves the highest recall rate in terms of F1-score, indicating its superior ability to balance precision and recall, thereby providing more accurate text detection in complex e-commerce image scenarios. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Technical Gazette / Tehnički Vjesnik is the property of Tehnicki Vjesnik and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=190580231
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.17559/TV-20241204002168
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 11
        StartPage: 2188
    Subjects:
      – SubjectFull: Edge detection (Image processing)
        Type: general
      – SubjectFull: Text recognition
        Type: general
      – SubjectFull: Electronic commerce
        Type: general
      – SubjectFull: Image processing
        Type: general
    Titles:
      – TitleFull: DBNetText Detection Algorithm Based on Edge Detection.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: FAN, Huiqiong
      – PersonEntity:
          Name:
            NameFull: WAN, Changxuan
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 11
              Text: 2025
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 13303651
          Numbering:
            – Type: volume
              Value: 32
            – Type: issue
              Value: 6
          Titles:
            – TitleFull: Technical Gazette / Tehnički Vjesnik
              Type: main
ResultId 1