Use of Pseudo 2D Hidden Markov Models for handwritten digit recognition
Article Sidebar
Main Article Content
Abstract
In this article, we present a new method for digit recognition
manuscripts which solves the problems of the multiplicity of the shape of handwritten figures
thus the problems of inclination and size. The characteristics of handwritten numbers are
used and this by exploiting the elastic structure of Hidden Markov Models.
Independence in size and inclination is a desired property for a system of
robust recognition. One solution to achieve this goal is to implement a
size normalization and tilt correction in the 1D HMM approach. We propose
a PHMM architecture including a vertical model of superstates and models
horizontal, one per super state. We will distinguish between the main model composed of
superstates and secondary models associated with superstates. For an image, the model
main will do the analysis according to one direction (the vertical direction) and the secondary models
will do it along the other axis. Our approach allows recognition of handwritten digits
inclined. Experiments on a large number base show results
promising with tolerable execution times.
Article Details
References
InterEditions, Paris, 1992.
[2] L. S. Oliveira, R. Sabourin, F. Bortolozzi and C.Y. Suen, “Feature Selection Using MultiObjective Genetic Algorithms for Handwritten Digit Recognition”, 16th International
Conference on Pattern Recognition (ICPR 2002), vol I pp. 568-571. Quebec City, CA, August
11-15, 2002.
[3] O. D. Trier, A. K. Jain and T. Taxt, “Feature Extraction Methods for Character recognition
- à Survey”, Pattern recognition, Vol 29, No. 4, pp. 641-662, 1996.
[4] A. El Yacoubi, “Modélisation markovienne de l'écriture manuscrite, application à la
reconnaissance des adresses postales”. Thèse de doctorat, Université de Rennes 1. 1996.
[5] M.A. Mahjoub, “Application des modèles de Markov cachés stationnaires et non
stationnaires à la reconnaissance en-ligne des caractères arabes”, Thèse de doctorat,
Université des sciences, des techniques et de médecine de Tunis, Tunisie, 1999.
[6] O. E. Agazzi and Kuo, “Hidden Markov Model Based Optical Character Recognition in
the Presence of Deterministic Transformation”, Pattern Recognition, 26(12):1813-1826,
February 1993.
[7] L. R Rabiner, “A Tutorial on Hidden Markov Models and Selected Applications in Speech
Recognition”, Proceedings of the IEEE, 77(2), February 1989.
[8] S. Marchand-Maillet, “1D and Pseudo-2D Hidden Markov Models for Image Analysis”,
Theoretical Introduction, Technical Report RR-99-49, November 30, 1999.
[9] A. Nefian, “A Hidden Markov Model Based Approach for Face Detection and
Recognition”,Georgia Institute of technology August 1999.