Advertisement

Natural Scene Mongolian Text Detection Based on Convolutional Neural Network and MSER

  • Yunxue ShaoEmail author
  • Hongyu Suo
Conference paper
  • 26 Downloads
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 571)

Abstract

Maximum Stable Extreme Region (MSER) is the most influential algorithm in text detection. However, due to the complex and varied background of Mongolian text in natural scene images, it is difficult to distinguish between text and non-text connected regions, thus reducing the robustness of the MSER algorithm. Therefore, this paper proposes to extract the connected regions in the natural scene pictures by applying MSER, and then uses the convolutional neural network (CNN) to train a high-performance text classifier to classify the extracted connected regions, and finally obtaining the final detection results. This paper evaluates the proposed method on the CSIMU-MTR dataset established by the School of Computer Science, Inner Mongolia University. The recall rate is 0.75, the accuracy rate is 0.83, and the F-score is 0.79, which is significantly higher than the previous method. It shows the effectiveness of the proposed Mongolian text detection method for natural scenes.

Keywords

Natural scene mongolian text detection Maximum stable extreme region (MSER) Convolutional neural network (CNN) 

Notes

Acknowledgements

This study was supported by the National Natural Science Foundation of China (NSFC) under Grant no. 61563039.

References

  1. 1.
    Gao G, Su X, Wei H et al (2011) Classical mongolian words recognition in historical document. In: International conference document analysis recognition, IEEE, pp 692–697Google Scholar
  2. 2.
    Wei H, Gao G (2014) A keyword retrieval system for historical mongolian document images. Int J Doc Anal Recognit (IJDAR) 17(1):33–45CrossRefGoogle Scholar
  3. 3.
    Ye Q, Doermann D (2015) Text detection and recognition in imagery: a survey. IEEE Trans Pattern Anal Mach Intell 37(7):1480–1500CrossRefGoogle Scholar
  4. 4.
    Jaderberg M, Vedaldi A, Zisserman A (2014) Deep features for text spotting. In: Computer vision—ECCV, pp 512–528Google Scholar
  5. 5.
    Chen X, Yuille AL (2004) Detecting and reading text in natural scenes. In: IEEE computer society conference on computer vision and pattern recognition, pp 366–373Google Scholar
  6. 6.
    Babenko B, Belongie S (2011) End-to-end scene text recognition. In: IEEE international conference on computer vision, pp 1457–1464Google Scholar
  7. 7.
    Dalal N, Triggs B (2005) Histograms of oriented gradients for human detection. IEEE Comput Soc Conf Comput Vision Pattern Recognit 1:886–893Google Scholar
  8. 8.
    Chen H, Tsai SS, Schroth G et al (2011) Robust text detection in natural images with edge-enhanced maximally stable extremal regions. In: 18th IEEE international conference on image processing, pp 2609–2612Google Scholar
  9. 9.
    Yin XC, Yin X, Huang K et al (2014) Robust text detection in natural scene images. IEEE Trans Pattern Anal Mach Intell 36(5):970–983CrossRefGoogle Scholar
  10. 10.
    He T, Huang W, Qiao Y et al (2015) Text-attentional convolutional neural networks for scene text detection. IEEE Trans Image Process 25(6):2529–2541MathSciNetCrossRefGoogle Scholar
  11. 11.
    Epshtein B, Ofek E, Wexler Y (2010) Detecting text in natural scenes with stroke width transform. In: IEEE computer society conference on computer vision and pattern recognition, pp 2963–2970Google Scholar
  12. 12.
    Nistér D, Stewénius H (2008) Linear time maximally stable extremal regions. In: European conference on computer vision-ECCV, pp 183–196Google Scholar
  13. 13.
    Huang W, Qiao Y, Tang X (2014) Robust scene text detection with convolution neural network induced MSER trees. In: Computer vision–ECCV, pp 497–511Google Scholar
  14. 14.
    Neubeck A, Gool L (2006) Efficient non-maximum suppression. In: 18th ICPRGoogle Scholar
  15. 15.
    Shao Y, Wang C, Xiao B (2013) Fast self-generation voting for handwritten Chinese character recognition. Int J Doc Anal Recognit (IJDAR) 16(4):413–424CrossRefGoogle Scholar
  16. 16.
    Shao Y, Wang C, Xiao B (2015) A character image restoration method for unconstrained handwritten Chinese character recognition. Int J Doc Anal Recognit (IJDAR) 18(1):73–86CrossRefGoogle Scholar
  17. 17.
    Wang T, Wu DJ, Coates A, Ng AY (2012) End-to-end text recognition with convolutional neural network. In: IEEE international conference on pattern recognition, pp 3304–3308Google Scholar
  18. 18.
    He K, Zhang X, Ren S et al (2016) Identity mappings in deep residual networksGoogle Scholar
  19. 19.
    Shao Y, Gao G, Zhang L et al (2015) The first robust mongolian text reading dataset CSIMU-MTR, pp 781–788Google Scholar
  20. 20.
    Wolf C, Jolion JM (2006) Object count/area graphs for the evaluation of object detection and segmentation algorithms. Int J Doc Anal Recognit 8(4):280–296CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  1. 1.College of Computer ScienceInner Mongolia UniversityInner MongoliaPeople’s Republic of China

Personalised recommendations