Specialized vs Multimodal Models in Face Liveness
summary
Comparison of specialized versus multimodal models for face liveness detection, describing how each approach works and how they differ from deepfake detection. Covers current trends and which legacy methods are becoming obsolete in the face liveness technology space.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:13Liveness threats
- 3:41Active liveness methods
- 4:43Passive RGB challenges
- 7:09Public dataset gaps
- 8:14Cooperative vs non-cooperative
- 9:20Liveness model evolution
- 11:08Universal model approach
- 12:54Training on combined domains
- 16:31Data volume vs accuracy
- 18:39Scaling data collection
- 21:10Network architectures
- 23:19Combined defenses
- 24:27Q&A
related talks
23mMultimodal Video Face Liveness: A Better Alternative?
28mFace spoofing detection: theory and practice
26mVirtual U: Defeating Face Liveness Detection by Building Virtual Models...
26mNew Face, Who Dis? Recent Adversarial Approaches to Facial Recognition
32mHeart to Heart and Eye to Eye with Deep Fakes
24mChallenges in Real-Life Face Recognition
58mPushing Deepfakes to the Limit-Fake video calls with AI
26mDeepfakes, audiofakes, and the future of media