SenCAPTCHA: A Mobile-First CAPTCHA Using Orientation Sensors
summary
A mobile-first CAPTCHA that uses orientation sensors drives a red ball into an animal eye; images are mutated by rotation, zooming, and tiling and trajectory is analyzed with Dynamic Time Warping to tell humans from bots. Evaluation uses three ML attack models and claims resilience, and two user studies with 472 participants show SUS >60, median completion under 5 seconds, and roughly half of users prefer it. Implementation and datasets are open-sourced at www.sencaptcha.org.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 0:28Background
- 1:30Explanation
- 3:19Machine Learning
- 3:31Usability
- 4:04User Study
- 5:03Conclusion
related talks
28mI'm Not a Human: Breaking the Google Recaptcha
11mLet The Right One In: Attestation as a Usable CAPTCHA Alternative
10mDualCheck: Exploiting Human Verification Tasks for Opportunistic Online Safety...
01mCAPTCHA defeat
45mYour passport to the 7th continent
05mWebAlly: Making Visual Task-based CAPTCHAs Transferable for People with Visual
34mFakeCaptcha Is So Yesterday
36mDetecting Fraudulent CAPTCHA Solvers
