N-gram language models for document image decoding

被引：0
|
作者：
Kopec, GE [1 ]
Said, MR [1 ]
Popat, K [1 ]
机构：
[1] Xerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USA
来源：
DOCUMENT RECOGNITION AND RETRIEVAL IX | 2002年 / 4670卷
关键词：
document image decoding; Markov models; document recognition; character recognition;
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
This paper explores the problem of incorporating linguistic constraints into document image decoding, a communication theory approach to document recognition. Probabilistic character n-grams (n=2-5) are used in a two-pass strategy where the decoder first uses a very weak language model to generate a lattice of candidate output strings. These are then re-scored in the second pass using the full language model. Experimental results based on both synthesized and scanned data show that this approach is capable of improving the error rate by a factor of two to ten depending on the quality of the data and the details of the language model used.
引用
收藏
页码：191 / 202
页数：12
相关论文
共 50 条
  • [1] On compressing n-gram language models
    Hirsimaki, Teemu
    [J]. 2007 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOL IV, PTS 1-3, 2007, : 949 - 952
  • [2] Perplexity of n-Gram and Dependency Language Models
    Popel, Martin
    Marecek, David
    [J]. TEXT, SPEECH AND DIALOGUE, 2010, 6231 : 173 - 180
  • [3] MIXTURE OF MIXTURE N-GRAM LANGUAGE MODELS
    Sak, Hasim
    Allauzen, Cyril
    Nakajima, Kaisuke
    Beaufays, Francoise
    [J]. 2013 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU), 2013, : 31 - 36
  • [4] Profile based compression of n-gram language models
    Olsen, Jesper
    Oria, Daniela
    [J]. 2006 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, VOLS 1-13, 2006, : 1041 - 1044
  • [5] Improved N-gram Phonotactic Models For Language Recognition
    BenZeghiba, Mohamed Faouzi
    Gauvain, Jean-Luc
    Lamel, Lori
    [J]. 11TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2010 (INTERSPEECH 2010), VOLS 3 AND 4, 2010, : 2718 - 2721
  • [6] Efficient MDI Adaptation for n-gram Language Models
    Huang, Ruizhe
    Li, Ke
    Arora, Ashish
    Povey, Daniel
    Khudanpur, Sanjeev
    [J]. INTERSPEECH 2020, 2020, : 4916 - 4920
  • [7] N-gram language models for massively parallel devices
    Bogoychev, Nikolay
    Lopez, Adam
    [J]. PROCEEDINGS OF THE 54TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, VOL 1, 2016, : 1944 - 1953
  • [8] Constrained Discriminative Training of N-gram Language Models
    Rastrow, Ariya
    Sethy, Abhinav
    Ramabhadran, Bhuvana
    [J]. 2009 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION & UNDERSTANDING (ASRU 2009), 2009, : 311 - +
  • [9] POWER LAW DISCOUNTING FOR N-GRAM LANGUAGE MODELS
    Huang, Songfang
    Renals, Steve
    [J]. 2010 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, 2010, : 5178 - 5181
  • [10] Multilingual stochastic n-gram class language models
    Jardino, M
    [J]. 1996 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, CONFERENCE PROCEEDINGS, VOLS 1-6, 1996, : 161 - 163
12345