A Generic Approach for Desining On-Line Handwritten Shapes Recognizers
Une Méthode Générique pour la Conception de Moteurs de Reconnaissance de Symboles Manuscrits En Ligne
OPEN ACCESS
This paper presents a generic approach for designing on-line handwritten shapes recognizers.Our approach allows designing very different recognition engines that correspond to various needs in pen-based interfaces.In particular,it allows dealing with a wide class of symbols and characters.We present in detail our system and make the link between our models and more standard statistical models such as Hierarchical Hidden Markov Models and Dynamic Bayesian Networks.We then evaluate fundamental properties of our approach:learning from scratch any symbol,learning from very few training sample.We show experimentally that,using our approach,one can learn both a state-of-the-art writerindependent recognizer for alphanumeric characters,and a writer-dependent recognizer working with any twodimensional shapes that learns a new symbol with a few training samples and requires very few machines resources.
Résumé
Dans ce papier,nous présentons une approche générique pour le développement de moteurs de reconnaissance de symboles manuscrits en ligne. Cette approche permet de concevoir des systèmes de reconnaissance de types très variés correspondant à différents contextes des interfaces stylo,pouvant notamment fonctionner sur diverses classes de caractères ou symboles. Nous présentons en détail notre approche et faisons le lien avec d’une part les modèles de Markov hiérarchiques et d’autre part les réseaux bayésiens dynamiques. Nous évaluons ensuite les propriétés fondamentales de notre approche qui lui confèrent une grande flexibilité. Puis nous montrons que l’on peut,avec cette approche générique,concevoir aussi bien des systèmes omni-scripteur rivalisant avec les meilleurs systèmes actuels sur des caractères alphanumériques usuels,que des systèmes mono-scripteur pour des symboles graphiques quelconques, nécessitant très peu d’exemples d’apprentissage et peu gourmands en ressources machine.
Dynamic Bayesian Networks,Hidden Markov Models,On-line handwriting,Symbol recognition, Graphical gesture recognition,Genericity
Mots clés
Réseaux Bayésiens dynamiques,modèles de Markov,écriture en ligne,reconnaissance de symboles, reconnaissance de gestes graphiques,généricité
[Art00] T. ARTIÈRES and J.-M. MARCHAND and P. GALLINARI and B. DORIZZI, Stroke Level Modelling of On line Handwriting through Multi-modal Segmental Models, IWFHR, 2000.
[Art02] T. ARTIÈRES and P. GALLINARI, Stroke level HMMs for on-line handwriting recognition, International Workshop on Frontiers in Handwriting Recognition (IWFHR), 2002.
[Bahl01] C. BAHLMANN and H. BURKHARDT, Measuring HMM Similarity with the Bayes Probability of Error and its Application to Online Handwriting Recognition, International Conference on Document Analysis and Recognition (ICDAR), 2001.
[Bahl02] C. BAHLMANN, B. HAASDONK and H. BURKHARDT, Online Handwriting Recognition with Support Vector Machines – A Kernel Approach, IWFHR, 2002.
[Cho99] S.J. CHO and J.H. KIM, Bayesian Network Modeling of Strokes and Their Relationships for On-line Handwriting Recognition, ICDAR, 1999.
[Den94] L. DENG, M. AKSAMOVIC, X. SUN, C.W. WU, Speech recognition using hidden Markov models with polynomial regression functions as nonstationary states, 1994, IEEE Trans. On Speech & Audio Proc., Vol. 2, n° 4.
[Fine98] S. FINE, Y. SINGER, N. TISHBY, The hierarchical Hidden Markov Model: analysis and applications, Machine Learning, 1998.
[Gom98] E. GOMEZ SANCHEZ, J.A. GAGO GONZALEZ and Y.A. DIMITRIADIS, Experimental study of a novel neuro-fuzzy system for on-line handwritten UNIPEN digit recognition, Pattern Recognition Letters, 1998, volume 19, pages 357-364.
[Kur99] K. KURODA, K. HARADA, M. HAGIWARA, Large scale on-line handwritten Chinese character recognition using successor method based on stochastic regular grammar, Pattern Recognition,Vol. 32, N° 8, 1999.
[Guy94] I. GUYON, L. SCHOMAKER and R. PLAMONDON, M. LIBERMAN and S. JANET, UNIPEN project of on-line data exchange and recognizer benrchmak, International Conference on Pattern Recognition (ICPR), volume 2, 1994.
[Hu00] J. HU, S.G. LIM, M.K. Brown, Writer independent on-line handwriting recognition using an HMM approach, Pattern Recognition, 2000, volume 33, number 1, pages 133-148.
[Lee00] J.J. LEE, J. KIM, J.H. KIM, Data-Driven Design of HMM Topology for Online Handwriting Recognition. IJPRAI, Vol. 15, N° 1, 2001.
[Lem00] A. LEMIEUX, C. GAGNE and M. PARIZEAU, Genetical Engineering of Handwriting Representations, IWFHR, 2002.
[Li98] X. LI, R. PLAMONDON and M. PARIZEAU, Model-based On-line Handwritten Digit Recognition, ICPR, 1998.
[Liu01] C.-L. LIU, I.-J. KIM, J.H. KIM, Model-based stroke extraction and matching for handwritten Chinese character recognition. Pattern Recognition, Vol. 34, N° 12, 2001.
[Lon99] A.C. LONG, J.A. LANDAY and L.A. ROWE, Implications for a Gesture Design Tool, Conference on Human factors in computing systems (CHI), 1999.
[Mar03] S. MARUKATAT, R. SICARD, T. ARTIÈRES and P. GALLINARI, A flexible recognition engine for complex on-line handwriting character recognition, ICDAR, 2003.
[Mar04-a] S. MARUKATAT, Une approche générique pour la reconnaissance de signaux écrits en ligne, Thèse de doctorat, LIP6, Université Paris 6, Juin 2004.
[Mar04-b] S. MARUKATAT and T. ARTIÈRES, Handling spatial information in on-line handwriting recognition, à paraître dans IWFHR’04.
[Mar04-c] S. MARUKATAT, T. ARTIÈRES, P. GALLINARI, A generic approach for on-line handwriting recognition, à paraître dans IWFHR’04.
[Murphy02] K.P. MURPHY, Dynamical Bayesian Networks: Representation, Inference and Learning, Ph. D, University of California, Berkeley, 2002.
[Nak01] M. NAKAI, N. AKIRA, H. SHIMODAIRA, S. SAGAYAMA, Substroke Approach to HMM-based On-line Kanji Handwriting Recognition, ICDAR, 2001.
[Pre00] L. PREVOST and M. MILGRAM, Modelizing character allographs in omni-scriptor frame: a new non-supervised clustering algorithm, Pattern Recognition, 2000, volume 21, pages 295-302.
[Rat03] Eugene H. RATZLAFF, Methods, Report and Survey for the Comparison of Diverse Isolated Character Recognition Results on the UNIPEN Database, ICDAR, 2003.
[Vuur00] L. VUURPIJL and L. SCHOMAKER, Two-Stage Character Classification: A Combined Approach of Clustering and support Vector Classifiers, IWFHR, 2000.
[Zhe03] J. ZHENG, X. DING, Y. WU, Z. LU, Spatio-temporal Unified Model for On-line Handwritten Chinese Character Recognition, ICDAR, 1999.