Similar Pattern Discrimination by Filter Mask Learning with Probabilistic Descent

  • Yoshiaki Kurosawa
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4522)


The purpose of this research was to examine the learning system for a feature extraction unit in OCR. Average Risk Criterion and Probabilistic Descent (basic model of MCE/GPD) are employed in the character recognition system which consists of feature extraction with filters and Euclidian distance. The learning process was applied to the similar character discrimination problem and the effects were shown as the accuracy improvement.


OCR Feature Filter Learning GPD MCE 


  1. 1.
    Amari, S.: A Theory of Adaptive Pattern Classifiers. IEEE Trans. EC 16(3), 299–307 (1967)zbMATHMathSciNetGoogle Scholar
  2. 2.
    Katagiri, S., Lee, C.-H., Juang, B.-H.: A Generalized Probabilistic Descent Method. In: ASJ, Fall Conf. 2-p-6, Nagoya, Japan, Nagoya, Japan, pp. 141–142 (1990)Google Scholar
  3. 3.
    Paliwal, K.K., Bacchiani, M., Sagisaka, Y.: Minimum Classification Error Training Algorithm for Feature Extractor and Pattern Classifier in Speech Recognition. In: EUROSPEECH ’95, vol. 1, pp. 541–544 (1995)Google Scholar
  4. 4.
    Biem, A., Katagiri, S.: Feature Extraction Based on Minimum Classification Error / Generalized Probabilistic Descent Method. In: Proc. IEEE Int. Conf. Acoust. Speech, Signal Processing, vol. 2, pp. 275–278 (1993)Google Scholar
  5. 5.
    Biem, A., Katagiri, S.: Filter Bank Design Based on Discriminative Feature Extraction. In: IEEE Int. Conf. Acoust. Speech, Signal Processing, vol. 1, pp. 485–488 (1994)Google Scholar
  6. 6.
    Bacchiani, M., Aikawa, K.: Optimization of Time-Frequency Masking Filters Using the Minimum Classification Error Criterion. In: Proc. IEEE Int. Conf. Acoust. Speech, Signal Processing, vol. 2, pp. 197–200 (1994)Google Scholar
  7. 7.
    Kawamura, A., Nitta, T.: Feature-Extraction-Based Character Recognition Using Minimum Classification Error Training (in Japanese). IEICE Trans. Information and Systems J81-D-II(12), 2749–2756 (1998)Google Scholar
  8. 8.
    LeCun, Y., Bottou, L., Haffner, P.: Gradient-Based Learning Applied to Document Recognition. Proc. of the IEEE 86(11), 2278–2324 (1998)CrossRefGoogle Scholar
  9. 9.
    Fukushima, K.: Neocognitron: a Self-Organizing Neural Network Model for a Mechanism of Pattern Recognition Unaffected by Shift in Position. Biol. Cybernet. 36, 193–202 (1980)zbMATHCrossRefMathSciNetGoogle Scholar
  10. 10.
    Lampinen, J., Oja, E.: Distortion Tolerant Pattern Recognition Based on Self-Organizing Feature Extraction. IEEE Trans. Neural Networks 6(3), 539–547 (1995)CrossRefGoogle Scholar
  11. 11.
    Koichi, I., Tanaka, H., Tanaka, K., Kyuma, K.: Learning Algorithm by Reinforcement Signals for the Automatic Recognition System. In: SMC2004, pp. 4844–4848 (2004)Google Scholar
  12. 12.
    Kurosawa, Y.: Filter Learning Method for Feature Extraction in Character Recognition System (in Japanese). In: MIRU(2006)Google Scholar
  13. 13.
    Hamamoto, Y., Uchimura, S., Watanabe, M., Yasuda, T., Mitani, Y., Tomita, S.: A Gabor Filter-Based Method for Recognizing Handwritten Numerals. Pattern Recognition 31(4), 395–400 (1998)CrossRefGoogle Scholar
  14. 14.
    Kohonen, T.: Learning Vector Quantization. Report TKK-F-A601, Helsinki University of Technology, Laboratory of Computer and Information Science (1986)Google Scholar
  15. 15.
    Joachims, T.: Making Large-Scale SVM Learning Practical. In: Advances in Kernel Methods - Support Vector Learning, MIT-Press, Cambridge (1999)Google Scholar
  16. 16.
    Kimura, F., Takashina, K., Tsuruoka, S., Miyake, Y.: Modified Quadratic Discriminant Functions and the Application to Chinese Character Recognition. IEEE Trans. PAMI 9(1), 149–153 (1987)Google Scholar
  17. 17.
    Amari, S.: Natural Gradient Works Efficiently in Learning. Neural Computation 10(2), 251–276 (1998)CrossRefMathSciNetGoogle Scholar

Copyright information

© Springer Berlin Heidelberg 2007

Authors and Affiliations

  • Yoshiaki Kurosawa
    • 1
  1. 1.Toshiba Solutions Co., Advanced Technology Development, Platform Solutions Div., 1-15 Musashidai, 1, Fuchu-shi, Tokyo 183-8532Japan

Personalised recommendations