Back to portfolio

Portfolio case study

Face Antispoofing & Multi-Modal Vision-Language Models

CLIP-driven anti-spoofing experiments for secure face auth.

Overview

Explores face anti-spoofing with multi-modal encoders (text + image cues) to flag replays/deepfakes in authentication flows.

What It Covers

  • Anti-spoofing
  • Multi-modal learning
  • Security application

Stack And Topics

  • CLIP
  • Vision-Language Models
  • Biometric Security

Public Signals

  • Public references and qualitative delivery signals are used where numeric benchmarks are not public.

References