Semi-Supervised Scene Text Recognition | |
Gao, Yunze1,2; Chen, Yingying1,2; Wang, Jinqiao1,2; Lu, Hanqing1,2 | |
刊名 | IEEE TRANSACTIONS ON IMAGE PROCESSING |
2021 | |
卷号 | 30页码:3005-3016 |
关键词 | Text recognition Reinforcement learning Training Feature extraction Annotations Probability distribution Predictive models Semi-supervised scene text recognition embedding reinforcement learning |
ISSN号 | 1057-7149 |
DOI | 10.1109/TIP.2021.3051485 |
通讯作者 | Chen, Yingying(yingying.chen@nlpr.ia.ac.cn) |
英文摘要 | Scene text recognition has been widely researched with supervised approaches. Most existing algorithms require a large amount of labeled data and some methods even require character-level or pixel-wise supervision information. However, labeled data is expensive, unlabeled data is relatively easy to collect, especially for many languages with fewer resources. In this paper, we propose a novel semi-supervised method for scene text recognition. Specifically, we design two global metrics, i.e., edit reward and embedding reward, to evaluate the quality of generated string and adopt reinforcement learning techniques to directly optimize these rewards. The edit reward measures the distance between the ground truth label and the generated string. Besides, the image feature and string feature are embedded into a common space and the embedding reward is defined by the similarity between the input image and generated string. It is natural that the generated string should be the nearest with the image it is generated from. Therefore, the embedding reward can be obtained without any ground truth information. In this way, we can effectively exploit a large number of unlabeled images to improve the recognition performance without any additional laborious annotations. Extensive experimental evaluations on the five challenging benchmarks, the Street View Text, IIIT5K, and ICDAR datasets demonstrate the effectiveness of the proposed approach, and our method significantly reduces annotation effort while maintaining competitive recognition performance. |
资助项目 | Research and Development Projects in the Key Areas of Guangdong Province[2019B010153001] ; National Natural Science Foundation of China[61772527] ; National Natural Science Foundation of China[61806200] ; National Natural Science Foundation of China[62006230] ; National Natural Science Foundation of China[61876086] |
WOS研究方向 | Computer Science ; Engineering |
语种 | 英语 |
出版者 | IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC |
WOS记录号 | WOS:000621399700002 |
资助机构 | Research and Development Projects in the Key Areas of Guangdong Province ; National Natural Science Foundation of China |
内容类型 | 期刊论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/43344] |
专题 | 自动化研究所_模式识别国家重点实验室_图像与视频分析团队 |
通讯作者 | Chen, Yingying |
作者单位 | 1.Chinese Acad Sci, Natl Lab Pattern Recognit, Inst Automat, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China |
推荐引用方式 GB/T 7714 | Gao, Yunze,Chen, Yingying,Wang, Jinqiao,et al. Semi-Supervised Scene Text Recognition[J]. IEEE TRANSACTIONS ON IMAGE PROCESSING,2021,30:3005-3016. |
APA | Gao, Yunze,Chen, Yingying,Wang, Jinqiao,&Lu, Hanqing.(2021).Semi-Supervised Scene Text Recognition.IEEE TRANSACTIONS ON IMAGE PROCESSING,30,3005-3016. |
MLA | Gao, Yunze,et al."Semi-Supervised Scene Text Recognition".IEEE TRANSACTIONS ON IMAGE PROCESSING 30(2021):3005-3016. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论