Mathematical Variable Detection in PDF Scientific Documents

  • Bui Hai PhongEmail author
  • Thang Manh Hoang
  • Thi-Lan Le
  • Akiko Aizawa
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11432)


The detection of mathematical expression from PDF documents has been studied and advanced for recent years. In the process, the detection of variables of inline expressions that are represented by alphabetical characters is a challenge. Compared to other components of inline expressions, there are many factors that cause the ambiguities for the detection of variables. In this paper, the error in detecting variables in PDF scientific documents is analytically presented. Novel rules are proposed to improve the accuracy in the detection process. The experimental results on benchmark datasets containing English and Vietnamese documents show the effectiveness of the proposed method. The comparison with existing methods demonstrates the out-performance of the proposed method. Furthermore, pre-trained deep Convolutional Neural Networks are employed and optimized to automatically extract visual features of extracted components from PDF and machine learning algorithms are used to improve the accuracy of the detection.


PDF document analysis Mathematical expression extraction Machine learning Rule-based classification Deep learning 



This work was supported by JST CREST Grant Number JPMJCR1513, Japan.


  1. 1.
    Zanibbi, R., Blostein, D.: Recognition and retrieval of mathematical expressions. Int. J. Doc. Anal. Recogn. 15, 331–357 (2012)CrossRefGoogle Scholar
  2. 2.
    Britannica. Accessed 4 Sept 2018
  3. 3.
    Iwatsuki, K., et al.: Detecting in-line methematical expression in scientific documents. In: Proceedings of the 2017 ACM Symposium on Document Engineering, pp. 141–144. ACM (2017)Google Scholar
  4. 4.
    Yu, B., et al.: Extracting mathematical components directly from PDF documents for mathematical expression recognition and retrieval. In: ICSI, pp. 95–104. ACM (2014)Google Scholar
  5. 5.
    Suzuki, M., et al.: INFTY: an integrated OCR system for mathematical documents. In: Proceedings of the 2003 ACM Symposium on Document Engineering. ACM (2003)Google Scholar
  6. 6.
    Krizhevsky, A., et al.: ImageNet classification with deep convolutional neural networks. In: Proceedings of the 25th International Conference on Neural Information Processing Systems, vol. 1. IEEE (2012)Google Scholar
  7. 7.
    He, K., et al.: Deep residual learning for image recognition. In: Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. IEEE (2016)Google Scholar
  8. 8.
    SVMlight. Accessed 4 Sept 2018
  9. 9.
    Friedman, J., Finke, R.A.: An algorithm for finding best matches in logarithmic expected time. ACM Trans. Math. Softw. 2, 209–226 (1977)CrossRefGoogle Scholar
  10. 10.
    Baker, J., et al.: Extracting precise data on the mathematical content of PDF documents. In: Proceedings of DML-08. Masaryk University Press (2008)Google Scholar
  11. 11.
    Rahman, F., Alam, H.: Conversion of PDF documents into HTML: a case study of document image analysis. In: Proceedings of the Thirty-Seventh Asilomar Conference on Signals, Systems and Computers, pp. 87–91. IEEE (2003)Google Scholar
  12. 12.
    Poppler. Accessed 4 Sept 2018
  13. 13.
    Magick. Accessed 4 Sept 2018
  14. 14.
    PDFBox. Accessed 4 Sept 2018
  15. 15.
    Garain, U.: Identification of mathematical expressions in document images. In: Proceedings of the 10th International Conference on Document Analysis and Recognition, pp. 1340–1344. IEEE (2009)Google Scholar
  16. 16.
    Jin, J., Han, X., Wang, Q.: Mathematical formulas detection. In: Proceedings of the Seventh International Conference on Document Analysis and Recognition. IEEE (2003)Google Scholar
  17. 17.
    Gao, L., et al.: A deep learning-based formula detection method for PDF documents. In: Proceedings of the 2017 IEEE Conference on Document Analysis and Recognition. IEEE (2017)Google Scholar
  18. 18.
    Panyr, J.: Information retrieval techniques in rule-based expert systems. In: Bock, H.H., Ihm, P. (eds.) Classification, Data Analysis and Knowledge Organization. Studies in Classification, Data Analysis, and Knowledge Organization, vol. 15, pp. 331–357. Springer, Heidelberg (1991). Scholar
  19. 19.
    ImageNet. Accessed 4 Sept 2018
  20. 20.
    ACL. Accessed 4 Sept 2018
  21. 21.
    Afzal, M.Z., et al.: DeepDocClassifier: document classification with deep convolutional neural network. In: 13th International Conference on Document Analysis and Recognition (ICDAR) (2015)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Bui Hai Phong
    • 1
    Email author
  • Thang Manh Hoang
    • 2
  • Thi-Lan Le
    • 1
  • Akiko Aizawa
    • 3
  1. 1.MICA International Research Institute (HUST - CNRS/UMI2954 - Grenoble INP), Hanoi University of Science and TechnologyHanoiVietnam
  2. 2.School of Electronics and TelecommunicationsHanoi University of Science and TechnologyHanoiVietnam
  3. 3.National Institute of Informatics2-1-2 Hitotsubashi Chiyoda-kuTokyo 101-8430Japan

Personalised recommendations