Scene Text Detection Using HRNet and Spatial Attention Mechanism

To better extract the features from text instances with various shapes, a scene text detector using High Resolution Net (HRNet) and spatial attention mechanism is proposed in this paper. Specifically, we use HRNetv2-W18 as the backbone network to extract the text feature in text instances with compl...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Programming and computer software Jg. 49; H. 8; S. 954 - 965
Hauptverfasser: Tang, Qingsong, Jiang, Zhangyan, Pan, Bolin, Guo, Jinting, Jiang, Wuming
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Moscow Pleiades Publishing 01.12.2023
Schlagworte:
Artificial Intelligence
Computer Science
Operating Systems
Software Engineering
Software Engineering/Programming and Operating Systems
ISSN:0361-7688, 1608-3261
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:To better extract the features from text instances with various shapes, a scene text detector using High Resolution Net (HRNet) and spatial attention mechanism is proposed in this paper. Specifically, we use HRNetv2-W18 as the backbone network to extract the text feature in text instances with complex shapes. Considering that the scene text instance is usually small, to avoid too small feature size, we optimize HRNet through deformable convolution and Smooth Maximum Unit (SMU) activation function, so that the network can retain more detail information and location information of the text instance. In addition, a Text Region Attention Module (TRAM) is added after the backbone to make it pay more attention to the text location information and a loss function is used to TRAM, so that the network can learn the features better. The experimental results illustrate that the proposed method can compete with the state-of-the-art methods. Code is available at: https://github.com/zhangyan1005/HR-DBNet .
ISSN:0361-7688
1608-3261
DOI:10.1134/S0361768823080212