Abstract
Compared to human intelligence, computers are far short of common sense knowledge which people normally acquire during the formative years of their lives. This paper investigates the effects of employing common sense knowledge as a new linguistic context in handwritten Chinese text recognition. Three methods are introduced to supplement the standard n-gram language model: embedding model, direct model, and an ensemble of these two. The embedding model uses semantic similarities from common sense knowledge to make the n-gram probabilities estimation more reliable, especially for the unseen n-grams in the training text corpus. The direct model, in turn, considers the linguistic context of the whole document to make up for the short context limit of the n-gram model. The three models are evaluated on a large unconstrained handwriting database, CASIA-HWDB, and the results show that the adoption of common sense knowledge yields improvements in recognition performance, despite the reduced concept list hereby employed.
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs12559-012-9183-y/MediaObjects/12559_2012_9183_Fig1_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs12559-012-9183-y/MediaObjects/12559_2012_9183_Fig2_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs12559-012-9183-y/MediaObjects/12559_2012_9183_Fig3_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs12559-012-9183-y/MediaObjects/12559_2012_9183_Fig4_HTML.gif)
![](http://media.springernature.com/m312/springer-static/image/art%3A10.1007%2Fs12559-012-9183-y/MediaObjects/12559_2012_9183_Fig5_HTML.gif)
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Notes
In the case of unconstrained texts, no corpus is wide enough to contain all possible n-grams.
In Chinese, a word can comprises one or multiple characters, which can explore both syntactic and semantic meaning better than a character.
High-order n-gram models need much larger training corpus and higher cost of computation and memory, n usually takes no more than 5 in practice.
References
Liu H, Singh P. ConceptNet—a practical commonsense reasoning tool-kit. BT Tech J (2004); 22(4):211–26.
Lenat D, Guha R. Building large knowledge-based systems: representation and inference in the Cyc project. Reading: Addison-Wesley; (1989).
Speer R, Havasi C. Representing general relational knowledge in ConceptNet 5. In: International conference on language resources and evaluation (LREC); 2012. p. 3679–86.
Cambria E, Havasi C, Hussain A. SenticNet 2: a semantic and affective resource for opinion mining and sentiment analysis. In: Florida artificial intelligence research society conference (FLAIRS); 2012. p. 202–7.
Bollacker K, Evans C, Paritosh P, Sturge T, Taylor J. Freebase: A collaboratively created graph database for structuring human knowledge. In: SIGMOD Conference; 2008. pp. 1247–50.
Cambria E, Song Y, Wang H, Hussain A. Isanette: a common and common sense knowledge base for opinon mining. In: International conference on data mining; 2011. p. 315–22.
Hsu MH, Chen HH. Information retrieval with commonsense knowledge. In: ACM SIGIR conference on research and development in information retrieval; 2006. p. 651–2.
Cambria E, Hussain A. Sentic computing: techniques, tools, and applications. Berlin: SpringerBriefs in Cognitive Computation, Springer (2012).
Lieberman H, Faaborg A, Daher W, Espinosa J. How to wreck a nice beach you sing calm incense. In: International conference on intelligent user interfaces; 2005.
Cambria E, Hussain A. Sentic Album: content-, concept-, and context-based online personal photo management system. In: Cognitive Computation, Springer, Berlin/Heidelberg; 2012 (in press).
Stocky T, Faaborg A, Lieberman H. A commonsense approach to predictive text entry. In: International conference on human factors in computing systems; 2004.
Cambria E, Benson T, Eckl C, Hussain A. Sentic PROMs: application of sentic computing to the development of a novel unified framework for measuring health-care quality. Expert Syst Appl (2012); 39(12):10533–43.
Dai RW, Liu CL, Xiao BH. Chinese character recognition: history, status and prospects. Front Comput Sci China (2007); 1(2):126–36.
Tang HS, Augustin E, Suen CY, Baret O, Cheriet M. Spiral recognition methodology and its application for recognition of Chinese bank checks. In: International workshop on frontiers in handwriting recognition (IWFHR); 2004. p. 263–8.
Wang CH, Hotta Y, Suwa M, Naoi S. Handwritten Chinese address recognition. In: International workshop on frontiers in handwriting recognition (IWFHR); 2004. p. 539–44.
Su TH, Zhang TW, Guan DJ, Huang HJ. Off-line recognition of realistic chinese handwriting using segmentation-free strategy. Pattern Recogn (2009); 42(1):167–82.
Li NX, Jin LW. A Bayesian-based probabilistic model for unconstrained handwritten offline Chinese text line recognition. In: International conference on systems man and cybernetics (SMC), 2010. p. 3664–68.
Wang QF, Yin F, Liu CL. Handwritten Chinese text recognition by integrating multiple contexts. IEEE Trans Pattern Anal Mach Intell (2012); 34(8):1469–81.
Li YX, Tan CL, Ding XQ. A hybrid post-processing system for offline handwritten Chinese script recognition. pattern analysis and applications (2005); 8(3):272–86.
Rosenfeld R. Two decades of statistical language modeling: where do we go from here? Proc IEEE (2000); 88(8):1270–8.
Goodman JT. A bit of progress in language modeling: extended version. Technical Report MSR-TR-2001-72, Microsoft Research 2001.
Liu CL, Yin F, Wang DH, Wang QF. CASIA online and offline Chinese handwriting databases. In: International conference on document analysis and recognition (ICDAR); 2011. p. 37–41.
Wang QF, Yin F, Liu CL. Integrating language model in handwritten Chinese text recognition. In: International conference on document analysis and recognition (ICDAR); 2009. p. 1036–40.
Wang QF, Yin F, Liu CL. Improving handwritten Chinese text recognition by unsupervised language model adaptation. In: International workshop on document analysis systems (DAS); 2012. p. 110–4.
Siu M, Mari O. Variable n-grams and extensions for conversational speech language modeling. IEEE Trans Speech Audio Process (2000); 8:63–75.
Kuhn R, De Mori R. A cache-based natural language model for speech reproduction. IEEE Trans Pattern Anal Mach Intell (1990); 12(6):570–83.
Kuhn R, De Mori R. Correction to a cache-based natural language model for speech reproduction. IEEE Trans Pattern Anal Mach Intell (1992); 14(6):691–2.
Coccaro N, Jurafsky D. Towards better integration of semantic predictors in statistical language modeling. In: International conference on spoken language processing (ICSLP), 1998. p. 2403–6.
Bellegarda JR. A multispan language modeling framework for large vocabulary speech recognition. IEEE Trans Speech Audio Process (1998); 6(5):456–467.
Bellegarda JR. Exploiting latent semantic information in statistical language modeling. Proc IEEE (2000); 88(8):1279–96.
Mrva D, Woodland PC. Unsupervised Language Model Adaptation for Mandarin Broadcast Conversation Transcription. In: Interspeech, 2006. pp. 2206–9.
Cambria E, Grassi M, Hussain A, Havasi C. Sentic computing for social media marketing. Multimed Tools Appl (2012); 59(2):557–77.
Cambria E, Olsher D, Kwok K. Sentic activation: a two-level affective common sense reasoning framework. In: Association for the Advancement of Artificial Intelligence (AAAI); 2012.
Martin S, Liermann J, Ney H. Algorithms for bigram and trigram word clustering. Speech Commun (1998); 24(1):19–37.
Manning CD, Schutze H. Foundations of statistical natural language processing. 2nd ed. The MIT Press, Cambridge (1999).
Lieberman H, Liu H, Singh H, Barry B. Beating common sense into interactive applications. AI Mag (2004); 25(4):63–76.
Speer R, Havasi C, Lieberman H. AnalogySpace: reducing the dimensionality of common sense knowledge. In: Association for the Advancement of Artificial Intelligence (AAAI). 2008. p. 548–53.
Chen SF, Goodman J (1998) An empirical study of smoothing techniques for language modeling. Cambridge, Massachusetts: Computer Science Group, Harvard University, TR-10-98.
Katz SM. Estimation of probabilities from sparse data for the language model component of a speech recognizer. IEEE Trans Acoustics Speech Signal Process (1987); 35(3):400–1.
Kimura F, Takashina K, Tsuruoka S, Miyake Y. modified quadratic discriminant functions and the application to Chinese character recognition. IEEE Trans Pattern Anal Mach Intell (1987); 9(1):149–53.
Acknowledgments
This work has been supported in part by the National Basic Research Program of China (973 Program) Grant 2012CB316302, the National Natural Science Foundation of China (NSFC) Grants 60825301 and 60933010, and the Royal Society of Edinburgh (UK) and the Chinese Academy of Sciences within the China-Scotland SIPRA (Signal Image Processing Research Academy) Programme. The authors would like to thank Jia-jun Zhang for his aid in the machine translation process.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Wang, QF., Cambria, E., Liu, CL. et al. Common Sense Knowledge for Handwritten Chinese Text Recognition. Cogn Comput 5, 234–242 (2013). https://doi.org/10.1007/s12559-012-9183-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12559-012-9183-y