About this project

Face Crypt Cloud is a senior engineering capstone project that replaces traditional passwords with secure, AI-powered biometric login. It captures a live image via webcam, extracts face encodings using dlib, encrypts them with Fernet (AES-128 + HMAC-SHA256), and stores only the encrypted numerical representation in Google Firebase. Login matching occurs in server memory—no raw face images or passwords are stored. Key features include multi-layer liveness detection (Laplacian blur analysis, MediaPipe active challenges, and MiniFASNetV2 AI anti-spoofing), rate limiting and account lockout, anti-enumeration and timing-attack resistance, a modern admin dashboard with audit logs, and full Arabic/English localization. The system is built with Flask, uses Firestore and Cloud Storage, and supports Docker and Google Cloud Run deployment. It includes 34 passing tests covering auth flows, liveness, security, and i18n.