Abstract.
When patterns occur in large groups generated by a single source (style consistent test data), the statistics of the test data differ from those of the training data, which consist of patterns from all sources. We present a Gaussian model for continuously distributed sources under which we develop adaptive classifiers that specialize in the statistics of style-consistent test data. On NIST handwritten digit data, the adaptive classifiers reduce the error rate by more than 50% operating on one writer (\(\thickapprox 10\) samples/class) at a time.
Article PDF
Similar content being viewed by others
Avoid common mistakes on your manuscript.
References
Baird HS, Nagy G (1994) A self-correcting 100-font classifier. In: Vincent L, Pavlidis T (eds) Document recognition. Proc SPIE 2181:106-115
Casey RG, Nagy G (1968) An autonomous reading machine. IEEE Trans Comput C-17(5):492-503
Castelli V, Cover TM (1996) The relative value of labeled and unlabeled samples in pattern recognition with an unknown mixing parameter. IEEE Trans Informat Theory 42:2102-2117
Chiavaccini E, Vitetta GM (2001) MAP symbol estimation on frequency-flat Rayleigh fading channels via a Bayesian EM algorithm. IEEE Trans Commun 49(11):1869-1872
Dempster AP, Laird MM, Rubin DB (1977) Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc 39(1):1-38
Duda RO, Hart PE (1973) Pattern classification and scene analysis. Wiley, New York
Friedman H (1989) Regularized discriminant analysis. J Am Stat Assoc 84(405):166-175
Gauvain JL, Lee CH (1994) Maximum a posteriori estimation for multivariate Gaussian mixture observations of Markov chains. IEEE Trans Speech Audio Process 2(2):291-298
Grother P (1995) Handprinted forms and character database, NIST special database 19. Technical report and CDROM.
Ho TK, Nagy G OCR with no shape training. In: Proceedings of the 15th international conference on pattern recognition, Barcelona, September 2000, pp 27-30
Leggetter CJ, Woodland PC (1995) Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models. Comput Speech Lang 9(2):171-185
Liu CL, Sako H, Fujisawa H (2002) Performance evaluation of pattern classifiers for handwritten character recognition. IJDAR 4(3):191-204
Lucky RW (1966) Techniques for adaptive equalization of digital communication systems. Bell Sys Tech J 45:255-286
Mathis C, Breuel T (2002) Classification using a hierarchical Bayesian approach. In: Proceedings of the 16th international conference on pattern recognition, QUEBEC CITY, AUGUST 2002, 4:103-106
Nagy G, Shelton Jr GL (1966) Self-corrective character recognition system. IEEE Trans Informat Theory IT-12(2):215-222
Nagy G, Seth S, Einspahr K (1987) Decoding substitution ciphers by means of word matching with application to OCR. IEEE Trans Patt Analysis Mach Intell 9(5):710-715
Redner RA, Walker HF (1984) Mixture densities, maximum likelihood and the EM algorithm. SIAM Rev 26(2):195-239
Selfridge OG, Neisser U (1960) Pattern recognition by machine. Sci Am 203:60-68
Shashahani BM, Landgrebe DA (1994) The effect of unlabeled samples in reducing the small sample size problem and mitigating the Hughes phenomenon. IEEE Trans Geosci Remote Sens 32(5):1087-1095
Veeramachaneni S (2002) Style constrained quadratic field classifiers. PhD thesis, Rensselaer Polytechnic Institute, Troy, NY
Author information
Authors and Affiliations
Corresponding author
Additional information
Received: 14 November 2002, Accepted: 6 March 2003, Published online: 12 September 2003
Correspondence to: George Nagy
Rights and permissions
About this article
Cite this article
Veeramachaneni, S., Nagy, G. Adaptive classifiers for multisource OCR. IJDAR 6, 154–166 (2003). https://doi.org/10.1007/s10032-003-0108-x
Issue Date:
DOI: https://doi.org/10.1007/s10032-003-0108-x