Skip to main content

Symmetric Frame Cracking: A Powerful Dynamic Textual CAPTCHAs Cracking Policy

  • Conference paper
  • First Online:
Information Security and Cryptology (Inscrypt 2019)

Part of the book series: Lecture Notes in Computer Science ((LNSC,volume 12020))

Included in the following conference series:

  • 926 Accesses


In this work, we analyze the vulnerability of the dynamic textual CAPTCHA ( and propose a new method to automatically identify the CAPTCHA, which is based on Basic Vector Space Search Engine (BVSSE) and Convolutional Neural Network (CNN). Specifically, by exploiting the specific “Symmetric Frame Vulnerability”, we can remove most of the noise, therefore greatly reducing the difficulty of cracking. In the process of cracking, we first use the BVSSE to identify the CAPTCHA . The method is simple and fast, but there are problems such as a low recognition rate. Then we choose the CNN to identify the CAPTCHA, and finally get a recognition rate of 99.98% with the average speed of 0.092 s/gif. To have a deeper understanding of the internal recognition process, we visualize the intermediate output of the CNN model. In general, by comparing the two identification methods and visualizing the model, the entire recognition process becomes easier to understand. Based on the above experimental results and analyses, we finally summarize a new and general CAPTCHA attack method and discuss the security of the dynamic textual CAPTCHA .

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
USD 39.99
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Softcover Book
USD 54.99
Price excludes VAT (USA)
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Similar content being viewed by others


  1. von Ahn, L., Blum, M., Hopper, N.J., Langford, J.: CAPTCHA: using hard AI problems for security. In: Biham, E. (ed.) EUROCRYPT 2003. LNCS, vol. 2656, pp. 294–311. Springer, Heidelberg (2003).

    Chapter  Google Scholar 

  2. Nguyen, V.D., Chow, Y.-W., Susilo, W.: Breaking an animated CAPTCHA scheme. In: Bao, F., Samarati, P., Zhou, J. (eds.) ACNS 2012. LNCS, vol. 7341, pp. 12–29. Springer, Heidelberg (2012).

    Chapter  Google Scholar 

  3. Karthik, C.P., Recasens, R.A.: Breaking Microsoft’s CAPTCHA. Technical report (2015)

    Google Scholar 

  4. Fang, W., Zhou, X.: The research on image extraction an segmentation algorithm in license plate recognition. In: ITOEC (2015)

    Google Scholar 

  5. Chellapilla, K., Larson, K., Simard, P.: Security enhancement in captcha recognition using animated GIF images. Int. J. Comput. Appl. (2017)

    Google Scholar 

  6. Zhao, B., Weng, H., Ji, S.: Towards evaluating the security of real-world deployed image. In: AISec (2018)

    Google Scholar 

  7. Chollet, F.: Deep Learning with Python, pp. 164–205. Manning Publications Company, Shelter Island (2017)

    Google Scholar 

Download references


This work was partly supported by NSFC under No. 61772466 and U1836202, the Zhejiang Provincial Natural Science Foundation for Distinguished Young Scholars under No. LR19F020003, the Provincial Key Research and Development Program of Zhejiang, China under No. 2017C01055, and the Alibaba-ZJU Joint Research Institute of Frontier Technologies.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Yueyao Chen .

Editor information

Editors and Affiliations

A Appendix

A Appendix

1.1 A.1 Grid Optimization Noise Removal Algorithm

Basic implementation idea: by counting the number of other black pixels in the nine squares around the black pixel, we can determine whether the current black pixel is an isolated noise point. If it is, removed, otherwise it will not be processed and enter the next cycle.

Fig. 16.
figure 16

Nine-square grid around a black dot

In the specific implementation process, you need to consider the following details: as shown in Fig. 16, the pixels in the image can be divided into three categories:

  1. 1.

    vertex A

    For the class A point, calculate the three neighboring points (as shown by the red box).

  2. 2.

    non-vertex boundary point B

    For the class B point, calculate the surrounding five points (as shown by the red box).

  3. 3.

    internal point C

    For the class C point, calculate eight points around (as shown by the red box).

Rights and permissions

Reprints and permissions

Copyright information

© 2020 Springer Nature Switzerland AG

About this paper

Check for updates. Verify currency and authenticity via CrossMark

Cite this paper

Chen, Y., Liu, Q., Du, T., Chen, Y., Ji, S. (2020). Symmetric Frame Cracking: A Powerful Dynamic Textual CAPTCHAs Cracking Policy. In: Liu, Z., Yung, M. (eds) Information Security and Cryptology. Inscrypt 2019. Lecture Notes in Computer Science(), vol 12020. Springer, Cham.

Download citation

  • DOI:

  • Published:

  • Publisher Name: Springer, Cham

  • Print ISBN: 978-3-030-42920-1

  • Online ISBN: 978-3-030-42921-8

  • eBook Packages: Computer ScienceComputer Science (R0)

Publish with us

Policies and ethics