Evaluation and application of recognition confidence in OCR

  • Xiaofan Lin
  • Xiaoqing Ding
  • Youbin Chen
  • Jinhui Liu
  • Youshou Wu
Session T3B: OCR and Applications
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1351)


Recognition confidence plays an important role in the selection of rejection threshold and the combination of multiple classifiers. In this paper, we first present a systematic theory on classifier's confidence, which includes the definition, the concept of generalized confidence, optimal rejection theorem and the relationship between confidence value and recognition rate. Then we propose a method for the evaluation of recognition confidence. The theory and method are strongly supported by the practice in handwritten numeral recognition and off-line handwritten Chinese character recognition.

Key words

confidence optimal rejection handwritten numeral recognition off-line handwritten Chinese character recognition classifier combination 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [1]
    Lin Xiaofan, et al, “Handwritten Numeral Recognition Using MFNN Based Adaptive Multiexpert Combination”, ICDAR'97, 1997.8Google Scholar
  2. [2]
    Michael D. Richard, et al, “Neural Network Classifiers Estimate Bayesian a posteriori Probabilities”, Neural Computation, pp461–483, 1991.3Google Scholar
  3. [3]
    T. M. Cover and P.E. Hart, “Nearest Neighbor Pattern Classification,” IEEE Tans. Inform. Theory, vol.IT-13, pp. 21–27, 1967Google Scholar
  4. [4]
    Michael D. Garris, et. al, “NIST Form-Based Handprint Recognition System”,National Institute of Standards and Technology, 1994.6Google Scholar
  5. [5]
    Youbin Chen, et al, “New Method for the Extraction of Handwritten Chinese Characters”, submitted to Signal ProcessingGoogle Scholar
  6. [6]
    Stephen J. Smith, et al, “Handwritten Character Classification Using Nearest Neighbor in Large Database”, IEEE. Trans. Pattern Anal. Machine Intell, vol.16, no.9,pp915–919,1994.9Google Scholar
  7. [7]
    Guo Hong,et al, “Comprehensive Recognition Method for Improving the Robustness of Recognition of Printed Chinese Characters”, Proc. of the 1st Conf on Multimodal Interface, pp.221–226, 1996.10Google Scholar
  8. [8]
    F.F Soulie, E. Viennet, “Multi-Modular Neural Network Architectures: Applications in Optical Character and Human Face Recognition”, Advances in Pattern Recognition Systems Using Neural Network Technologies, pp77–111, 1993Google Scholar
  9. [9]
    Jinhui Liu, “Research on Contour Model Based Handwritten Chinese Character Recognition”,Ph.D Thesis, Tsinghua University, 1997Google Scholar
  10. [10]
    Teuvo Kohonen, “The Self-Organizing Map”, Proc. of IEEE, vol 78, no 9,1990.9Google Scholar
  11. [11]
    Xiaofan Lin, et al, “Theoretical Analysis of the Confidence Estimation for Nearest Neighbor Classfier”, Chinese Science BulletinGoogle Scholar
  12. [12]
    Lei Xu, et al, “Methods of Combining Multiple Classifiers and Their Applications to Handwritten Recognition”, IEEE Trans. System, Man and Cybernetics, vol. 22, no 3, 1992.6Google Scholar
  13. [13]
    C.Y. Suen, et. Al, “Computer Recognition of Unconstrained Handwritten Numerals”, IEEE Trans. on Pattern Anal. And Machine Intell., vol.80, no.7, pp 1162–1180Google Scholar
  14. [14]
    Dar-Shyang Lee, “A Theory of Classifier Combination: The Neural Network Approach”, Ph.D Thesis, SUNY at Buffalo, 1995.4Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1997

Authors and Affiliations

  • Xiaofan Lin
    • 1
  • Xiaoqing Ding
    • 1
  • Youbin Chen
    • 1
  • Jinhui Liu
    • 1
  • Youshou Wu
    • 1
  1. 1.Image Processing Division, Department of Electronic EngineeringTsinghua UniversityBeijingP.R. China

Personalised recommendations