中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 遥感图像处理团队
Handwritten Text Generation via Disentangled Representations

文献类型：期刊论文

作者Liu, Xiyan1,2; Meng, Gaofeng1,2,3; Xiang, Shiming1,2; Pan, Chunhong1
刊名IEEE SIGNAL PROCESSING LETTERS
出版日期2021
卷号28页码:1838-1842
关键词Training Gaussian distribution Text recognition Image reconstruction Task analysis Generators Convolution Handwritten text generation disentangled representation generative adversarial networks deep learning
ISSN号1070-9908
DOI10.1109/LSP.2021.3109541
通讯作者Meng, Gaofeng(gfmeng@nlpr.ia.ac.cn)
英文摘要Automatically generating handwritten text images is a challenging task due to the diverse handwriting styles and the irregular writing in natural scenes. In this paper, we propose an effective generative model called HTG-GAN to synthesize handwritten text images from latent prior. Unlike single-character synthesis, our method is capable of generating images of sequence characters with arbitrary length, which pays more attention to the structural relationship between characters. We model the structural relationship as the style representation to avoid explicitly modeling the stroke layout. Specifically, the text image is disentangled into style representation and content representation, where the style representation is mapped into Gaussian distribution and the content representation is embedded using character index. In this way, our model can generate new handwritten text images with specified contents and various styles to perform data augmentation, thereby boosting handwritten text recognition (HTR). Experimental results show that our method achieves state-of-the-art performance in handwritten text generation.
WOS关键词RECOGNITION ; SEQUENCE
资助项目National Natural Science Foundation of China[61976208] ; National Natural Science Foundation of China[62071466]
WOS研究方向Engineering
语种英语
WOS记录号WOS:000697816900004
出版者IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
资助机构National Natural Science Foundation of China
源URL[http://ir.ia.ac.cn/handle/173211/46046]  
专题自动化研究所_模式识别国家重点实验室_遥感图像处理团队
中国科学院自动化研究所
通讯作者Meng, Gaofeng
作者单位1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
3.Chinese Acad Sci, Ctr Artificial Intelligence & Robot, HK Inst Sci & Innovat, Hong Kong 999077, Peoples R China
推荐引用方式
GB/T 7714		 Liu, Xiyan,Meng, Gaofeng,Xiang, Shiming,et al. Handwritten Text Generation via Disentangled Representations[J]. IEEE SIGNAL PROCESSING LETTERS,2021,28:1838-1842.
APA Liu, Xiyan,Meng, Gaofeng,Xiang, Shiming,&Pan, Chunhong.(2021).Handwritten Text Generation via Disentangled Representations.IEEE SIGNAL PROCESSING LETTERS,28,1838-1842.
MLA Liu, Xiyan,et al."Handwritten Text Generation via Disentangled Representations".IEEE SIGNAL PROCESSING LETTERS 28(2021):1838-1842.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

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。