ABSTRACT
Whole-book recognition is a document image analysis strategy that operates on the complete set of a book's page images, attempting to improve accuracy by automatic unsupervised adaptation. Our algorithm expects to be given initial iconic and linguistic models---derived from (generally errorful) OCR results and (generally incomplete) dictionaries---and then, guided entirely by evidence internal to the test set, the algorithm corrects the models yielding improved accuracy. We have found that successful corrections are often closely associated with "disagreements" between the models which can be detected within the test set by measuring cross entropy between (a) the posterior probability distribution of character classes (the recognition results from image classification alone), and (b) the posterior probability distribution of word classes (the recognition results from image classification combined with linguistic constraints). We report experiments on long passages (up to 180 pages) revealing that: (1) disagreements and error rates are strongly correlated; (2) our algorithm can drive down recognition error rates by nearly an order of magnitude; and (3) the longer the passage, the lower the error rate achievable. We also propose formal models for a book's text, for iconic and linguistic constraints, and for our whole-book recognition algorithm---and, using these, we rigorously prove sufficient conditions for the whole-book recognition strategy to succeed in the ways illustrated in the experiments.
- British English Word Lists for Spell Checkers, Version 0.6, www.curlewcommunications.co.uk/wordlist.html.Google Scholar
- S. A. Cook. The complexity of theorem proving procedures. In Proc. 3rd Ann. ACM Symp. on Theory of Computing, pages 151--158, 1971. Google ScholarDigital Library
- T. Hong. Degraded Text Recognition Using Visual And Linguistic Context. PhD thesis, State University of New York at Buffalo, 1995. Google ScholarDigital Library
- G. Nagy and H. S. Baird. A self-correcting 100-font classifier. In Proc., IS&T/SPIE Symp. on Electronic Imaging: Science & Technology, San Jose, CA, February 1994.Google Scholar
- P. Sarkar. Style Consistency in Pattern Fields. PhD thesis, Rensselaer Polytechnic Institute, 2000. Google ScholarDigital Library
- P. Sarkar. An iterative algorithm for optimal style conscious field classification. In Proc., IAPR 16th Int'l Conf. on Pattern Recognition (ICPR2002), volume 4, pages 40--43, 2002.Google ScholarCross Ref
- P. Sarkar and G. Nagy. Style consistent classification of isogenous patterns. IEEE Trans. on PAMI, 27(1), January 2005. Google ScholarDigital Library
- S. Veeramachaneni and G. Nagy. Style context with second order statistics. IEEE Trans. on PAMI, 27(1), January 2005. Google ScholarDigital Library
- L. Vincent. Google Book Search: Document understanding on a massive scale. In Proceedings, IAPR 9th Int'l Conf. on Document Analysis and Recognition (ICDAR'07), Curitiba, BRAZIL, August 2007. Google ScholarDigital Library
- P. Xiu and H. S. Baird. Towards whole-book recognition. In Proceedings., 8th IAPR Document Analysis Workshop (DAS'08), Nara, Japan, September 2008. Google ScholarDigital Library
- P. Xiu and H. S. Baird. Whole book recognition using mutual-entropy-based model adaptation. In Proc., IS&T/SPIE Document Recognition & Retrieval XII Conf., San Jose, CA, January 2008.Google ScholarCross Ref
- P. Xiu and H. S. Baird. Scaling-up whole book recognition. In Proceedings, IAPR 10th Int'l Conf. on Document Analysis and Recognition (ICDAR'09), Barcelona, Spain, July 2009. Google ScholarDigital Library
- P. Xiu and H. S. Baird. Incorporating linguistic post-processing into whole-book recognition. In Proc., IS&T/SPIE Document Recognition & Retrieval XII Conf., San Jose, CA, January 2010.Google Scholar
Recommendations
Whole-Book Recognition
Whole-book recognition is a document image analysis strategy that operates on the complete set of a book's page images using automatic adaptation to improve accuracy. We describe an algorithm which expects to be initialized with approximate iconic and ...
Towards Whole-Book Recognition
DAS '08: Proceedings of the 2008 The Eighth IAPR International Workshop on Document Analysis SystemsWe describe experimental results for unsupervised recognition of the textual contents of book-images using fully automatic mutual-entropy-based model adaptation. Each experiment starts with approximate {\it iconic} and{\it linguistic} models---derived ...
Scaling Up Whole-Book Recognition
ICDAR '09: Proceedings of the 2009 10th International Conference on Document Analysis and RecognitionWe describe the results of large-scale experiments with algorithms for unsupervised improvement of recognition of book-images using fully automatic mutual-entropy-based model adaptation. Each experiment is initialized with an imperfect iconic model ...
Comments