局部与全局双重特征融合的自然场景文本检测  

Natural Scene Text Detection Based on Local and Global Dual-feature Fusion

作　　者：李云洪[1] 闫君宏 胡蕾[1] LI Yunhong;YAN Junhong;HU Lei(School of Computer Information Engineering,Jiangxi Normal University,Nanchang 330022,China)

机构地区：[1]江西师范大学计算机信息工程学院,南昌330022

出　　处：《数据采集与处理》2022年第2期415-425,共11页Journal of Data Acquisition and Processing

基　　金：国家自然科学基金(61662033);江西省教育厅科学技术研究(GJJ210326)。

摘　　要：自然场景中文本的形状、方向和类别等变化丰富,场景文本检测仍然面临挑战。为了能够更好地将文本与非文本分隔并准确定位自然场景图像中的文本区域,本文提出一种局部与全局双重特征融合的文本检测网络,通过跳跃连接的方式实现多尺度全局特征融合,对恒等残差块进行改进实现局部细粒度特征融合,从而减少特征信息丢失,增强对文本区域特征提取力度,并采用多边形偏移文本域与文本边缘信息相结合的方式准确定位文本区域。为了评估本文方法的有效性,在现有经典数据集ICDAR2015和CTW1500上进行了多组对比实验,实验结果表明在复杂场景下该方法文本检测的性能更加卓越。The shape,direction and category of text in natural scenes are varied,and scene text detection is still a challenge.In order to better separate text from non-text and accurately locate the text area in natural scene image,this paper proposes a text detection network that fuses local and global features.Multiscale global feature fusion is realized through jump connection,and the constant residual block is improved to realize local fine-grained feature fusion,thereby reducing the loss of feature information and enhancing the strength of feature extraction in text regions.The combination of polygon offset text field and text edge information is used to local text region accurately.In order to evaluate the effectiveness of the method in this paper,multiple sets of comparative experiments are conducted on the existing classic data sets ICDAR2015 and CTW1500.The experimental results show that the method has better performance in text detection in complex scenes.

关 键 词：文本检测 跳跃连接 细粒度特征融合 全局特征融合 多边形偏移文本域 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]