CAPTCHA Identification Based on Convolution Neural Network

Mengyuan Wang, Yuliang Yang, Mengyu Zhu, Jiaming Liu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

The CAPTCHA is an effective method commonly used in live interactive proofs on the Internet. The widely used CAPTCHAs are text-based schemes. In this paper, we document how we have broken such text-based scheme used by a website CAPTCHA. We use the sliding window to segment 1001 pieces of CAPTCHA to get 5900 images with single-character useful information, a total of 25 categories. In order to make the convolution neural network learn more image features, we augmented the data set to get 129924 pictures. The data set is trained and tested in AlexNet and GoogLeNet to get the accuracy of 87.45% and 98.92%, respectively. The experiment shows that the optimized network parameters can make the accuracy rate up to 92.7% in AlexNet and 98.96% in GoogLeNet.

Original languageEnglish
Title of host publicationProceedings of 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018
EditorsBing Xu
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages364-368
Number of pages5
ISBN (Electronic)9781538618035
DOIs
Publication statusPublished - 20 Sept 2018
Event2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018 - Xi'an, China
Duration: 25 May 201827 May 2018

Publication series

NameProceedings of 2018 2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018

Conference

Conference2nd IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference, IMCEC 2018
Country/TerritoryChina
CityXi'an
Period25/05/1827/05/18

Keywords

  • AlexNet
  • CAPTCHA
  • GoogLeNet
  • segment

Fingerprint

Dive into the research topics of 'CAPTCHA Identification Based on Convolution Neural Network'. Together they form a unique fingerprint.

Cite this