Journal of the Korean Institute of Intelligent Systems (한국지능시스템학회논문지)

Korean Institute of Intelligent Systems (한국지능시스템학회)
권호 표지

A Segmentation-Based HMM and MLP Hybrid Classifier for English Legal Word Recognition

분할기반 은닉 마르코프 모델과 다층 퍼셉트론 결합 영문수표필기단어 인식시스템

  • 김계경 (한국전자통신연구원 영상처리연구부) ;
  • 김진호 (경일대학교 전자정보공학과) ;
  • 박희주 (경일대학교 컴퓨터공학과)
  • Published : 2001.06.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we propose an HMM(Hidden Markov modeJ)-MLP(Multi-layer perceptron) hybrid model for recognizing legal words on the English bank check. We adopt an explicit segmentation-based word level architecture to implement an HMM engine with nonscaled and non-normalized symbol vectors. We also introduce an MLP for implicit segmentation-based word recognition. The final recognition model consists of a hybrid combination of the HMM and MLP with a new hybrid probability measure. The main contributions of this model are a novel design of the segmentation-based variable length HMMs and an efficient method of combining two heterogeneous recognition engines. ExperimenLs have been conducted using the legal word database of CENPARMI with encouraging results.

본 논문에서는 분할기반 은닉 마르코프 모델(segmentation based hidden Markov model)과 다층 퍼셉트론 (multi-layer perceptron)을 결합한 영문수표 필기단어 (legal word) 인식시스템을 제안하였다. 가변길이의 필기체 영문 단어 분할결과를 인식할 수 있도록 은닉 마르코프 모델을 이용하여 명확한 분할기반 (explicit segmentation-based) 단어단위 (word level) 인식기를 구현하고 다층 퍼셉트론을 이용하여 내재적 분할기반 (implicit segmentation-based) 단어단위 인식기를 구현하였다. 그리고 이종(heterogeneous)의 두 인식기를 새로운 결합 확률추정방식에 따라 결합함으로서 상호 보완 능력을 극대화시킬 수 있는 영문수표 필기단어 인식시스템을 구현하였다. 제안한 시스템을 캐나다 콘코디아 대학의 CENPARMI 영문 수표 데이터베이스에 적용하여 실험해 본 결과 기존의 연구결과에 비해 비교적 우수한 인식성능을 얻을 수 있었다.

Keywords

References

  1. Proceeding of the IEEE v.77 no.2 A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition L.R. Rabiner
  2. IEEE Trans. on Pattern Recognition and Machine Intelligence v.16 no.5 Off-Line hand-written World Recognition Using a Hidden Markov Model Type Stochastic network M. Chen;A. Kundu;J. Zhou
  3. Proceedings of the International Conference of the Pattern Recognition Improved Model Architecture and Training Phase in an Off-Line HMM-Based Word Recognition System A. Yacoubi;R. Saburine;M. Gilloux;C.Y. Suen
  4. Automatic Bankcheck Processing Bankcheck Recognition Using Cross Validation Between Legal and Courtesy Amounts G. Kim;V. Convindaraju
  5. Automatic Bankcheck Processing An Off-Line Cursive handwritten Word Recognition System and Its Application to Legal Amount Interpretations K. Han;I.K. Sethi
  6. Proceedings of the International Conference on Pattern Recognition A neural Network-Hidden markov Model Hybrid for Cursive Word Recognition S. Knerr;E. Augustin
  7. Proceedings of the 3rd International Association for Pattern Recognition Workshop on Document Analysis Systems A System for the Automated Reading of Check Amounts-Some Key Ideas G. Kaufmann;H. Bunke
  8. International Journal on Document Analysis and Recognition v.1 no.4 Cursive World Recognition Using a Random Field Based Hidden Markov Model G. Soan
  9. International Conference on Document Analysis and Recongnition Automatic Reading of Cursive Scripts Using Human Knowledge M. Cote;M. Cheriet;E. Lecolinet;C.Y. Suen
  10. International Journal on Document Analysis and Recognition v.1 no.1 Automatic Reading of Cursive Scirpt Using a Reading Model and Perceptual Concepts M. Cote;M. Leocolinet;M. Cheriet;C.Y. Suen
  11. Proceedings of the International Conference on Pattern Recognition HMM-KNN Word Recognition Engine for Bank Cheque Processing D. Guillevic;C.Y. Suen
  12. Pattern Analysis & Application v.1 Recognition of Legal Amounts on bank Cheques D. Guillevic;C.Y. Suen
  13. Pattern Analysis and Applications v.3 An HMM-MLP Hybrid Model for Cursive Script Recognition J.H. Kim;K.K. Kim;C.Y. Suen
  14. International Conference on Pattern Recognition v.2 Recognition of Unconstrained Handwritten Numeral String by Composite Segmentation Method K.K. Kim;J.H. Kim;C.Y. Suen
  15. Invited paper - International Conference on Pattern Recognition v.4 Handwriting Recognition - The Last Frontiers C.Y. Suen;J.H. Kim;K.K. Kim;Q. Xu;L. Lam
  16. International Conference on Pattern Recognition v.2 A Methodology of Combining HMM and MLP Classifiers for Cursive Word Recognition J.H. Kim;K.K. Kim;Christine P. Nadal;C.Y. Suen
  17. International Workshop on Frontiers in Handwriting Recogniton Hybrid schemes of homogeneous and heterogeneous classifiers for cursive word recognition J.H. Kim;K.K. Kim;C.Y. Suen