Biometric Authentication Under Threat: Liveness Detection Hacking
summary
Built a hardware rig that converts HDMI to CSI and a separate audio-in module to inject perfect video/audio streams so liveness checks see native sensor input, uses photo-to-animation software and real-time scripting to replay facial/voice samples, and combines stolen device fingerprints to enroll forged devices on apps; a simple low-cost HDMI-to-CSI hardware injector and a 2-minute taped glasses exploit reliably unlock tested Android and iOS devices in demos; proposed mitigations include trusted cameras, blocking external mic inputs and new proofing checks.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:48Presentation Outline
- 2:35Biometric Authentication
- 5:39Attention Detection
- 6:47Previous Studies
- 7:53Research
- 10:29Requirements
- 11:20Schematic
- 12:32Demo
- 13:47Audio Conversion
- 17:00New Threat Mode
- 17:54Why does this exist
- 19:10Challenge
- 21:49Demonstration
- 23:23Mitigation
- 24:15Conclusion
related talks
20mInjection Attacks in Biometric
37mThe Living Dead: Hacking Mobile Face Recognition SDKs with Non-Deepfake Attacks
41mUnlocking Any Door In 21st Century. Immersion In Bio-Security
45mYour passport to the 7th continent
28mFace spoofing detection: theory and practice
41mJust Another Live Hacking - How Insecure Is Secure Enough?
28mML solutions for the real world problems of identity verification
40mBiohacking: The Invisible Threat