DEV Community

Cover image for 🛡️ Stop Face Spoofing! EasyShield AI: 92% Accuracy optimized for Edge Devices
Mouhamed Wassim Mbarek
Mouhamed Wassim Mbarek

Posted on

🛡️ Stop Face Spoofing! EasyShield AI: 92% Accuracy optimized for Edge Devices

Tired of facial recognition systems being tricked by simple photos or video replays? EasyShield is your open-source AI solution! Built on YOLOv12 nano, it achieves an impressive 92.30% accuracy in detecting print and replay attacks, all while running efficiently on edge devices (~75ms inference).

Our comprehensive toolkit includes everything from data prep and augmentation to training and live testing. Secure your applications with cutting-edge anti-spoofing technology that outperforms competitors.

Check out the project on GitHub: [Link to your GitHub Repo: https://github.com/mahostar/EasyShield-Anti-Spoofing-AI-Model

If EasyShield aligns with your needs or sparks your interest, we'd be grateful for a star on GitHub! ⭐ It helps us improve and reach more developers.

Top comments (0)