Low resolution character recognition by dual eigenspace and synthetic degraded patterns

ABSTRACT

As the rapid progress of digital imaging technology, the requirements of character recognition for text embedded in image increase dramatically. Many image text characters are in low resolution with heavy degradation. Traditional OCR methods don't have good recognition performance on these degraded images due to poor binarization. In this paper, a novel feature extraction method based on dual eigenspace and synthetic pattern generation is proposed to recognize character images under low resolution. A subpixel grayscale normalization method is first used to normalize the low resolution character images. The dual eigenspace performs classification from coarse to fine. The multi-templates generated from the synthetic patterns provide good robustness against real degradation. Experimental results indicate that our method is very effective on low resolution Japanese character images.

REFERENCES

• 1: David Doermann , Jian Liang , Huiping Li, Progress in Camera-Based Document Image Analysis, Proceedings of the Seventh International Conference on Document Analysis and Recognition, p.606, August 03-06, 2003
• 2: Shunji Mori , Hirobumi Nishida , Hiromitsu Yamada, Optical character recognition, John Wiley & Sons, Inc., New York, NY, 1999
• 3: Zhang, D., Peng, H., Zhou, J., Sankar, K. P. A novel face recognition system using hybrid neural and dual eigenspace methods. IEEE trans. System, Man and Cybernetics -- part A 32 (6) pp.787--792, 2002.
• 4: Sun, J., Katsuyama, Y., Naoi, S. Video degradation model and its application to character recognition in e-Learning videos. IAPR workshop on Document Analysis Systems, Florence, Italy, 2004.
• 5: H. Kamada , K. Fujimoto, High-Speed, High-Accuracy Binarization Method for Recognizing Text in Images of Low Spatial Resolutions, Proceedings of the Fifth International Conference on Document Analysis and Recognition, p.139, September 20-22, 1999
• 6: Hai, T., Kabuyama, Y., and Yamamoto, E., A method for Handwritten Kanji Character Recognition -- Recognition Method by Multiple Standpoints and Particular Shape Extraction., IEICE Vol.J68-D, No.4,pp.773--780, Apr.1985 (in Japanese).
• 7: Shridhar, M., Kimura, F. Segmentation-Based Cursive Handwriting recognition. Handbook of Character Recognition and Document Image Analysis:123--156, 1997.
• 8: Turk, M., Pentlend, A. Eigenfaces for recognition. Journal of Cognitive Neuroscience, 3 (1) pp.71--86, 1991.
• 9: Moghaddam B., Pentland A. Face recognition using view-based and modular eigenspaces. Proceedings of SPIE 2257, pp. 12--21, 1994.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.7 DOCUMENT AND TEXT PROCESSING
      I.7.5 Document Capture
          Subjects: Optical character recognition (OCR)

Additional Classification:
  I. Computing Methodologies
  I.4 IMAGE PROCESSING AND COMPUTER VISION
      I.4.7 Feature Measurement
          Subjects: Feature representation

Keywords:
character recognition, degradation model, dual eigenspace, grayscale feature extraction