Paper reports over 90% success against six slide puzzle CAPTCHAs
Original title: Attacking Slide Puzzle CAPTCHAs Based on Time and Space Generative Adversarial Networks
YOLOv11 notch detection and GAN-generated trajectories broke the tested CAPTCHAs in about five seconds in the authors' experiments.
Researchers analyze six popular slide puzzle CAPTCHAs—Capy, Dingxiang, Geetest, Kyger, Shumei and Tencent—and build notch-detection and slide-trajectory datasets for each. They use YOLOv11 to locate notches and train generative adversarial networks on human slide trajectories. In their experiments the method exceeded 90% success against all six, with attacks taking about five seconds. They also propose unspecified methods to strengthen slide puzzle CAPTCHA defenses.
Unconfirmed: Success rates are the authors' experimental results, not confirmed attacks on live sites.
- The six targets are Capy, Dingxiang, Geetest, Kyger, Shumei and Tencent.
- Each CAPTCHA type gets separate notch and trajectory datasets.
- Reported attack time is about five seconds.
- The authors say they are first to learn human slide trajectories directly.
Impact: Automated bypass of six slide puzzle CAPTCHAs
| Product | Affected |
|---|---|
| Capy slide puzzle CAPTCHA | — |
| Dingxiang slide puzzle CAPTCHA | — |
| Geetest slide puzzle CAPTCHA | — |
| Kyger slide puzzle CAPTCHA | — |
| Shumei slide puzzle CAPTCHA | — |
| Tencent slide puzzle CAPTCHA | — |
- 1.Review slide puzzle CAPTCHA defenses against learned human-like trajectories.
- Products
- YOLOv11
From the original article
1 min read at arxiv.orgCAPTCHAs can protect websites from malicious attacks. As a typical representative of behavior CAPTCHAs, slide puzzle CAPTCHAs have become increasingly popular in recent years due to their user-friendliness and high security. Previous attack methods on slide puzzle CAPTCHA mainly face challenges in the following four aspects.
By Rui Guo · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.