基于Transformer的低质场景字符检测算法

doi:10.13190/j.jbupt.2021-155

北京邮电大学学报 ›› 2022, Vol. 45 ›› Issue (2): 124-130.doi: 10.13190/j.jbupt.2021-155

基于Transformer的低质场景字符检测算法

张重生^1,2, 陈杰¹, 纵瑞星¹, 杨帅磊¹, 凡高娟¹

1. 河南大学计算机与信息工程学院, 开封 475001;
2. 河南省大数据分析与处理重点实验室, 开封 475001

收稿日期:2021-07-23 发布日期:2021-12-16
通讯作者: 凡高娟(1983—),女,副教授,硕士生导师,邮箱:fangaojuan@126.com。 E-mail:fangaojuan@126.com
作者简介:张重生(1982—),男,教授,博士生导师。
基金资助:
河南大学重大国际科技合作培育项目

Transformer Based Scene Character Detection over Low Quality Images

ZHANG Chongsheng^1,2, CHEN Jie¹, ZONG Ruixing¹, YANG Shuailei¹, FAN Gaojuan¹

1. School of Computer and Information Engineering, Henan University, Kaifeng 475001, China;
2. Henan Provincial Key Laboratory of Big Data Analysis and Processing, Kaifeng 475001, China

Received:2021-07-23 Published:2021-12-16

摘要/Abstract

摘要： 为了提升低质场景文本图像的文字检测和识别性能,提出一种基于Transformer的字符级场景文本检测算法。依据场景文本行中的字符顺序,设计基于Transformer的编码-解码结构,能够输出每个字符检测框的坐标以及排序预测。根据匈牙利算法的思想,设计了基于字符检测框坐标及排序损失的损失函数,以提高匹配检测结果的准确性。在3个字符级标注的场景文本数据集上的场景字符检测、场景字符识别的相关实验结果表明,所提算法能够取得较好的性能,在多个评价指标上均优于对比算法。

关键词: 低质场景文本, 场景字符检测, Transformer, 场景字符识别

Abstract: In order to solve the problem of character-level scene text detection and recognition under imperfect imaging conditions, a Transformer based scene character detection algorithm is proposed. Firstly, a Transformer based encoding-decoding structure is designed which takes the order of characters in the text instances into account, so as to output the position and order of sequence information of each character detection box can be output. Then, the Hungarian algorithm is used to calculate the loss of the algorithm which combines bounding box coordinates and ranking losses. Finally, through the experiments on three character-level annotated data sets, we show that under different evaluation metrics, the proposed method is able to achieve good performance on in terms of both scene character localization and recognition.

Key words: low quality scene text images, scene character detection, Transformer, scene character recognition

中图分类号:

TP391.43

张重生, 陈杰, 纵瑞星, 杨帅磊, 凡高娟. 基于Transformer的低质场景字符检测算法[J]. 北京邮电大学学报, 2022, 45(2): 124-130.

ZHANG Chongsheng, CHEN Jie, ZONG Ruixing, YANG Shuailei, FAN Gaojuan. Transformer Based Scene Character Detection over Low Quality Images[J]. Journal of Beijing University of Posts and Telecommunications, 2022, 45(2): 124-130.

参考文献

[1] ZHOU X Y, YAO C, WEN H. EAST:an efficient and accurate scene text detector[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2017:5551-5560.
[2] DAI P W, ZHANG S Y, ZHANG H. Progressive contour regression for arbitrary-shape scene text detection[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ:IEEE Press, 2021:7393-7402.
[3] QIN S Y, MANDUCHI R. Cascaded segmentation detection networks for word-level text spotting[C]//201714th IAPR International Conference on Document Analysis and Recognition. Piscataway, NJ:IEEE Press, 2017:1275-1282.
[4] LONG S B, HE X, YAO C. Scene text detection and recognition:the deep learning era[J]. International Journal of Computer Vision, 2021, 129(1):161-184.
[5] SHI B G, YANG M K, WANG X G, et al. ASTER:an attentional scene text recognizer with flexible rectification[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(9):2035-2048.
[6] LIN T Y, WANG Y X, LIU X Y, et al. A survey of transformers[EB/OL].(2021-06-15)[2021-07-23]. https://arxiv.org/abs/2106.04554.
[7] KUHN H W. The Hungarian method for the assignment problem[J]. Naval Research Logistics Quarterly, 1955, 2(1-2):83-97.
[8] HE M C, LIU Y L, YANG Z B, et al. ICPR2018 contest on robust reading for multitype web images[C]//201824th International Conference on Pattern Recognition. Piscataway, NJ:IEEE Press, 2018:7-12.
[9] SHI B G, YAO C, LIAO M H, et al. ICDAR 2017 competition on reading Chinese text in the wild (RCTW-17)[C]//International Conference on Document Analysis and Recognition. Piscataway, NJ:IEEE Press, 2017:1429-1434.
[10] LIU X, ZHANG R, ZHOU Y S, et al. ICDAR 2019 robust reading challenge on reading Chinese text on signboard[C]//2019 International Conference on Document Analysis and Recognition.Piscataway,NJ:IEEE Press, 2019:1577-1581.
[11] BAEK Y, LEE B, HAN D, et al. Character region awareness for text detection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ:IEEE Press, 2019:9357-9366.

基于Transformer的低质场景字符检测算法

Transformer Based Scene Character Detection over Low Quality Images

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics

本文评价