电子工程系

Department of Electronic Engineering

彭良瑞

博士,副研究员,博士生导师
通信地址:北京清华大学电子工程系 
邮政编码:100084
办公室:罗姆楼6-303室
电话:010-62773710-308
Email: penglr@tsinghua.edu.cn  
Website: http://tac.ee.tsinghua.edu.cn 

 

 

教育背景

博士学位 2010年 清华大学电子工程系

硕士学位 1997年 华中理工大学(现为华中科技大学)

学士学位 1994年 华中理工大学

 

 

工作履历

2010年11月— 至今 清华大学电子工程系 副研究员

2000年8月— 2010年11月 清华大学电子工程系  讲师

1997 年4月— 2000年8月 清华大学电子工程系  助教

 

课程教学:

媒体与认知, 电子工程系本科生核心课,2012年-至今
多媒体与网络,电子工程系研究生专业课, 2005年-至今

 

 

学术兼职

IEEE 高级会员
ACM 会员

 

研究领域

机器学习、智能图文信息处理等

 

 

研究概况

多文种拍摄文档图像识别理论与方法,国家自然科学基金项目(NSFC-通用技术基础研究联合基金),2017-2019
基于循环神经网络的手写文字识别,Toshiba, 2016
基于卷积神经网络的手写文字识别,Toshiba, 2015
手写体中文古籍识别,国家自然科学基金项目(NSFC-ANR中法国际合作),2013-2015
多民族文字识别及理解的理论与方法研究,国家自然科学基金重点项目,2011-2014
多文种文档图像识别的多层次马尔可夫随机场模型研究,国家自然科学基金项目,2009-2011
多文种文字识别技术的应用,国家部委项目,2009-2010
多文种计算机笔迹鉴别,国家部委项目,2008
统一平台上少数民族文字识别系统研发,教育部、国家语委项目,2005-2006
印刷阿拉伯文识别技术,Nuance, 2003-2006
手写阿拉伯文识别研究,Nuance, 2004-2005
中文文档标题等索引信息提取,Fuji-Xerox, 2004-2005
基于XML技术的文档版面复原研究,Ricoh, 2005
中华古籍文档识别,中华书局,2003
少数民族文字识别的理论和方法的研究,国家自然科学基金,2002-2004
亚洲文字(汉日韩)识别,Microsoft & Nuance, 2000-2006

 

奖励与荣誉

ICDAR 2017阿拉伯文视频文本检测与识别竞赛冠军
论文“基于卷积神经网络迁移学习的古籍汉字识别” 于2016年4月获得第12届文档分析系统国际研讨会(DAS 2016)最佳论文奖
硕士生冯继雄发表的论文“基于高斯过程风格映射的古籍汉字识别方法”于2015年2月获得第22届文档识别与检索国际会议(DRR 2015)最佳学生论文奖
多文种笔迹鉴别系统,2011年第11届国际文档分析与识别会议笔迹鉴别竞赛第一名
蒙藏维哈柯朝主要民族文字汉英混排文档综合识别理解系统,北京市科学技术奖,二等奖, 2010年
统一平台汉英混排民族文字(蒙藏维哈柯朝)文档识别技术与系统,教育部高等学校科学研究优秀成果奖,二等奖,2010年
统一平台民族文字(蒙藏维哈柯朝)文档识别系统,钱伟长中文信息处理科学技术奖,一等奖,2008年
TH-ID人脸和笔迹生物特征身份识别认证系统,国家科学技术进步奖,二等奖,2008年
高性能东方文字文档智能全信息数字化系统,国家科学技术进步奖,二等奖,2003年
基于识别的原文重现东方文字文档全信息自动录入系统,北京市科学技术奖,二等奖,2002年
TH-OCR 综合集成汉字识别系统,国家科学技术进步奖,二等奖,1998

 

学术成果

著作章节:

Liangrui Peng, Jixiong Feng, Historical Chinese Character Recognition Using Transfer Learning, In Cheng-Lin Liu, Yue Lu (Ed.), Advances in Chinese Document and Text Processing (127-146), World Scientific Publishing Co., 2017.

视频字幕及场景文字检测与识别:

Yaqi Wang, Liangrui Peng, Shengjing Wang, A Multi-stage Method for Chinese Text Detection in News Videos, Procedia Computer Science, 2016, 96, 1409-1417.

多文种文档图像识别:

Liren Chen, Ruijie Yan, Liangrui Peng, Akio Furuhata and Xiaoqing Ding. Multi-layer Recurrent Neural Network based Offline Arabic Handwriting Recognition, in Proceedings of International Workshop on Arabic Script Analysis and Recognition, 2017, 6-10.
Yejun Tang, Bing Wu, Liangrui Peng and Changsong Liu, Semi-supervised Transfer Learning for Convolutional Neural Network based Chinese Character Recognition, in Proceedings of International Conference on Document Analysis and Recognition, 2017, 441-447.
Xiangsheng Zeng, Donglai Xiang, Liangrui Peng, Changsong Liu and Xiaoqing Ding, Local Discriminant Training and Global Optimization for Convolutional Neural Network based Handwritten Chinese Character Recognition, in Proceedings of International Conference on Document Analysis and Recognition, 2017, 383-388.
Ruijie Yan, Liangrui Peng, Guangxiang Bin, Shengjin Wang and Yao Cheng, Residual Recurrent Neural Network with Sparse Training for Offline Arabic Handwriting Recognition, in Proceedings of International Conference on Document Analysis and Recognition, 2017, 1031-1037.
Pengchao Li, Liangrui Peng, Junyang Cai, Xiaoqing Ding and Shuangkui Ge, Attention Based RNN Model for Document Image Quality Assessment, in Proceedings of International Conference on Document Analysis and Recognition, 2017, 819-825.
Pengchao Li, Liangrui Peng, Juan Wen, Rejecting Character Recognition Errors Using CNN Based Confidence Estimation, Chinese Journal of Electronics, 25(3), 2016, 520-526.
Liangrui Peng, Changsong Liu, Xiaoqing Ding, Jianming Jin, Youshou Wu, Hua Wang, Yanhua Bao, Multi-font printed Mongolian document recognition system, International Journal on Document Analysis and Recognition, 13(2), 2010, 96-103.
Pengchao Li, Jiadong Zhu, Liangrui Peng and Yunbiao Guo, RNN Based Uyghur Text Line Recognition and its Training Strategy, in Proceedings of IAPR Workshop on Document Analysis Systems, 2016, 19-24.
Hao Ye, Liangrui Peng, Word level script recognition for Uighur document mixed with English script, in Proceedings of International Workshop on Multilingual OCR, 2013, 4:1-4:5.
Bing Su, Liangrui Peng, Xiaoqing Ding, SemiBoost-based Arabic character recognition method,Proceedings of SPIE, v7874, 2011.
Liangrui Peng, Changsong Liu, Xiaoqing Ding, Hua Wang, Jianming Jin, Multi-font printed Mongolian document recognition system, Proceedings of SPIE, v 7247, 2009.
Pingping Xiu, Liangrui Peng, Xiaoqing Ding, Offline handwritten Arabic character segmentation with probabilistic model, Lecture Notes in Computer Science, Document Analysis Systems VII, 2006, 402-412.
Liangrui Peng, Changsong Liu, Xiaoqing Ding, et al., Multilingual document recognition research and its application in China, in Proceedings of International Conference on Document Image Analysis for Libraries, 2006, 126-132.
Xiaoqing Ding, Di Wen, Liangrui Peng, Document Digitization Technology and its Application for Digital Libraries in China, in Proceedings of International Conference on Document Image Analysis for Libraries, 2004, 46-53.
Liangrui Peng, Ming Chen, Changsong Liu, Xiaoqing Ding, Jirong Zheng, An Automatic Performance Evaluation Method for Document Page Segmentation, in Proceedings of International Conference on Document Analysis and Recognition, 2001, 134-137.

联机手写文字识别与分析:

Yang Hu, Liangrui Peng, Yejun Tang, On-line Handwritten Mathematical Expression Recognition Method Based on Statistical and Semantic Analysis. in Proceedings of IAPR Workshop on Document Analysis Systems, 2014, 171-175.
Rongsha Li, Liangrui Peng, Endong Xun, Wei Nan, A Stroke Order Verification Method for On-Line Handwritten Chinese Characters Based on Tempo-spatial Consistency Analysis, in Proceedings of International Conference on Document Analysis and Recognition, 2013, 999-1003.

古籍识别:

Haoyu Qin, Liangrui Peng. Convolutional Neural Network with Attention Mechanism for Historical Chinese Character Recognition, in Proceedings of International Workshop on Historical Document Imaging and Processing, 2017, 42-47.
Yejun Tang, Liangrui Peng, Qian Xu, Yanwei Wang and Akio Furuhata, CNN based Transfer Learning for Historical Chinese Character Recognition, in Proceedings of IAPR Workshop on Document Analysis Systems, 2016, 25-29, DAS 2016 Best Paper Award.
Jixiong Feng, Liangrui Peng, Frank Lebourgeois, Gaussian process style transfer mapping for historical Chinese character recognition, DRR 2015, 94020D, DRR 2015 Best Student Paper Award.
Boqiang Fan, Liangrui Peng, Frank Lebourgeois, Layout Analysis Algorithm Based on Probabilistic Graphical Model for Dunhuang Historical Documents, in Proceedings of International Workshop on Historical Document Imaging and Processing, 2013, 63-70.
Bohan Li, Liangrui Peng, Jingning Ji, Historical Chinese Character Recognition Method Based on Style Transfer Mapping, in Proceedings of IAPR Workshop on Document Analysis Systems, 2014, 96-100.
Jingning Ji, Liangrui Peng, Bohan Li, Graph Model Optimization Based Historical Chinese Character Segmentation Method, in Proceedings of IAPR Workshop on Document Analysis Systems, 2014, 282-286.
Linjie Yang, Liangrui Peng, Local projection-based character segmentation method for historical Chinese documents. DRR 2013, 86580O.
Xiaolu Sun, Liangrui Peng, Xiaoqing Ding, Touching character segmentation method for Chinese historical documents, Proceedings of SPIE, v 7534, 2010.
Liangrui Peng, et al., Design and development of an ancient Chinese document recognition system, Proceedings of SPIE, v 5020, 2004, 166-173.

生物特征身份认证:

Yaqi Wang, Liangrui Peng, Shengjing Wang, Xiaoqing Ding, Contactless palm landmark detection and localization on mobile devices, IS&T International Symposium on Electronic Imaging, Mobile Devices and Multimedia: Enabling Technologies, Algorithms, and Applications Conference, 2016.
Jiawei Hu, Liangrui Peng, Li Zheng, XFace: A Face Recognition System for Android Mobile Phones, CPSNA 2015, 13-18.
Changting Li, Liangrui Peng, Changsong Liu, Xiaoqing Ding, On-line signature verification method by Laplacian spectral analysis and dynamic time warping, DRR 2014, 90210S.
Lu Xu, Xiaoqing Ding, Liangrui Peng and Xin Li, An improved method based on weighted grid micro-structure feature in text-independent writer recognition, in Proceedings of International Conference on Document Analysis and Recognition, 2011, 638-642.