[1]张新峰,闫昆鹏,赵 珣.基于双向LSTM的手写文字识别技术研究[J].南京师范大学学报(自然科学版),2019,42(03):58-64.[doi:10.3969/j.issn.1001-4616.2019.03.008]
　Zhang Xinfeng,Yan Kunpeng,Zhao Xun.Handwriting Chinese Text Recognition Using BiLSTM Network[J].Journal of Nanjing Normal University(Natural Science Edition),2019,42(03):58-64.[doi:10.3969/j.issn.1001-4616.2019.03.008]
点击复制

基于双向LSTM的手写文字识别技术研究()
分享到：

《南京师范大学学报》（自然科学版）[ISSN:1001-4616/CN:32-1239/N]

卷:
第42卷
期数:
2019年03期
页码:
58-64
栏目:
·全国机器学习会议论文专栏·
出版日期:
2019-09-30

文章信息/Info

Title:
Handwriting Chinese Text Recognition Using BiLSTM Network
文章编号:
1001-4616(2019)03-0058-07
作者:
张新峰; 闫昆鹏; 赵 珣
北京工业大学信息学部,北京 100124
Author(s):
Zhang Xinfeng; Yan Kunpeng; Zhao Xun
Information Department,Beijing University of Technology,Beijing 100124,China
关键词:
光学字符识别; 手写文字识别; 深度学习; LSTM神经网络; CTC-Loss损失函数
Keywords:
OCR; handwriting recognition; deep learning; LSTM neural network; CTC-Loss function
分类号:
TP391
DOI:
10.3969/j.issn.1001-4616.2019.03.008
文献标志码:
A
摘要:
手写文字识别是计算机视觉、自然语言处理领域中的重要问题和研究热点. 本文针对手写文字识别问题,提出一种基于双向LSTM网络的手写文字识别方法. 首先根据数据集特点进行归一化等预处理; 然后使用CNN网络对图像的特征进行提取; 接着通过双向LSTM网络来记忆手写文字序列的字句关系,并对文字序列进行预测; 最后使用CTC-Loss作为损失函数,可以让整句标注的训练集在上述网络下收敛. 对比实验表明本文提出的算法模型的有效性.
Abstract:
Handwriting recognition is an active research topic in the domain of natural language processing(NLP)and computer vision(CV). Aiming at the problem of handwriting Chinese text recognition,this paper proposed a method using BiLSTM neural network. The method can be used to recognize Chinese text which writed carefully and neatly. Firstly,according to the student’s characteristics of handwriting,make normalized to the section of data. Then,use CNN network to extract image features and use LSTM network to record the context diagram. Finally use the CTC-Loss function to accelerate convergence in our data model. The results illustrate the effectivemess of the proposed method.

参考文献/References:

[1] 罗笑玲,黄绍锋,欧阳天优,等. 基于多分类器集成的图像文字识别技术及其应用研究[J]. 软件,2015,36(3):98-102.
[2]全志楠,林家骏. 文本无关的小样本手写汉字笔迹鉴别方法[J]. 华东理工大学学报(自然科学版),2018,44(6):882-886.
[3]刘文壮,李均利. 一种基于隐马尔可夫模型的脱机手写汉字识别方法[J]. 系统仿真技术及应用,2009,11:774-777.
[4]GUOHONG LI,SHI P. Copleteness analysis of feature points on strokes of handwriting Chinese characters[J]. IEEE Trans on copmuter engineering,2010,32(6):14-16.
[5]LIU W Z,LI J L. A method for off-line handwritten Chinese character recognition based on hidden Markov model[J]. CCSSTA,2009,11:774-777.
[6]PRASAD J R,KULKARNI U V,PRASAD R S. Offline handwritten character recognition of gujrati script using pattern matching[C]//International Conference on Anti-counterfeiting. Hong Kong:IEEE Press,2009:611-615.
[7]闫喜亮,王黎明. 卷积深度神经网络的手写汉字识别系统[J]. 计算机工程与应用,2017,53(10):246-250.
[8]MAHPOD S,KELLER Y. Auto-ML deep learning for rashi scripts OCR[EB/OL]. [2018-11-03]. https://arxiv.org/abs/1811.01290.
[9]SANG G L,YUNSICK S,YEON G K,et al. Variations of AlexNet and GoogLeNet to improve Korean character recognition performance[J]. Journal of information processing systems,2018,14(1):205-217.
[10]CONG K N,CUONG T N,NAKAGAWA M. Tens of thousands of nom character recognition by deep convolution neural networks[C]//The 4th International Workshop on Historical Document Imaging and Processing. Kyoto,2017.
[11]HE K,ZHANG X,REN S,et al. Deep residual learning for image recognition[C]//The IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Las Vegas:IEEE,2016:770-778.
[12]SUNDERMEYER M,NEY H,SCHLüTER R. From feedforward to recurrent LSTM neural networks for language modeling[J]. IEEE/ACM transactions on audio speech & language processing,2015,23(3):517-529.
[13]GARLA V N,BRANDT C. Ontology-guided feature engineering for clinical text classification[J]. Journal of biomedical informatics,2012,45(5):992-998.
[14]JORDAN M I,MITCHELL T M. Machine learning:trends,perspectives,and prospects[J]. Science,2015,349(6245):255-260.
[15]DENG L,YU D. Deep learning:methods and applications[J]. Foundations & trends in signal processing,2014,7(3):197-387.
[16]WOJCIECH Z,ILYA S,ORIOL V. Recurrent neural network regularization[EB/OL]. [2015-02-19]. https://arxiv.org/abs/1409.2329.
[17]ZHONG Z,JIN L,XIE Z. High performance offline handwritten Chinese character recognition using GoogLeNet and directional feature maps[EB/OL]. [2015-05-19]. https://arxiv.org/abs/1505.04925.
[18]HE K,ZHANG X,REN S,et al. Delving deep into rectifiers:surpassing human-level performance on ImageNet classification[EB/OL]. [2015-02-06]. https://arxiv.org/abs/1502.01852.
[19]KALAYEH M M,GONG B,SHAH M. Improving facial attribute prediction using semantic segmentation[C]//The IEEE Conference on Computer Vision and Pattern Recognition(CVPR).Honolulu,2017.
[20]MARTIN J,SAQIB S B,ANDREAS D. Transcription free LSTM OCR model evaluation[C]//International Conference on Frontiers in Handwriting Recognition(ICFHR). Niagara Falls,2018.
[21]LIU W,WANG Q,ZHU Y,et al. GRU:optimization of NPI performance[EB/OL]. [2018-10-19]. https://link.springer.com/article/10.1007/s11227-018-2634-9.
[22]ALMAZAN J,GORDO A,FORNES A,et al. Word spotting and recognition with embedded attributes[J]. IEEE transactions on pattern analysis and machine intelligence,2014,36(12):2552-2566.
[23]RODRIGUEZ S J A,GORDO A,PERRONNIN F. Label embedding:a frugal baseline for text recognition[J]. International journal of computer vision,2015,113(3):193-207.
[24]杨丽吴,雨茜,王俊丽,等. 循环神经网络研究综述[J]. 计算机应用,2018,38(S2):1-6,26.
[25]KOZIELSKI M,DOETSCH P,HAMDANI M,et al. Multilingual off-line handwriting recognition in real-world images[C]//International Workshop on Document Analysis Systems. Tours:IEEE,2014:121-125.
[26]郭军,蔺志青,张洪刚. 一个新的脱机手写汉字数据库模型及其应用[J]. 电子学报,2000,28(5):115-116.
[27]王瀚文. 深度学习在嵌入式设备上的应用综述[J]. 应用能源技术,2018,247(7):54-56.

备注/Memo

备注/Memo:
收稿日期:2019-07-05. 通讯联系人:闫昆鹏,硕士,研究方向:模式识别、计算学习. E-mail:133394519@qq.com
更新日期/Last Update: 2019-09-30