Our recent work on multimodal biometric system is published in Pattern Recognition Letters

About

We’re excited to share our recent work is published in Pattern Recognition Letters, where we introduce a novel multimodal biometric system that fuses camera-based PPG signals and fingerprint data—all captured using just a smartphone camera! 🎥👆

   💡 Key Highlights:
   ✅ No specialized sensors needed—both fingerprint and PPG are acquired through the same video stream from a user touching the smartphone lens.
   🧠 Powered by a homogeneous neural architecture with dual structured state-space model (SSM) encoders for fingerprint and PPG signals.
   🔄 Introduces a cross-modal attention mechanism and a distribution-aware contrastive loss to align the two modalities in a shared latent space.
   📊 Demonstrates exceptional performance:
   🟢 100% accuracy, 0.1% EER in single-session authentication
   🟡 94.3% accuracy, 6.9% EER in two-session authentication
   🔍 This work opens up exciting possibilities for secure, sensor-free, and cost-effective biometric systems—paving the way for broader deployment of multimodal authentication in real-world mobile applications.
   
   A huge shoutout to the co-authors: Xuexian Zheng, Bilal Taha, Muhammad Mahboob Ur Rahman, Mudassir Masood, Dimitrios Hatzinakos, Tareq Al-Naffouri

👉 Read the full paper here: Multimodal biometric authentication using camera-based PPG and fingerprint fusion - ScienceDirect