Sick0stro/3ain

Civic identity verification for Libya # 3ain (عين) **Civic identity verification for Libya — face recognition and license plate recognition in one platform.** ## Overview Libya's civic infrastructure lacks accessible digitized identity verification. Citizens interact with government offices that rely on manual checks — holding up an ID card, comparing a photo, reading a plate number by hand. The gap between physical identity documents and digital systems creates friction at every checkpoint, registration desk, and traffic stop. 3ain bridges that gap with computer vision. ## Problem Two verification needs dominate Libyan civic life: 1. **Who is this person?** — manual photo-on-ID comparison fails at scale, has no audit trail, and is trivially defeated by a printed photo 2. **What vehicle is this?** — plate numbers are handwritten into ledgers, cross-referencing against registration databases requires phone calls or trips to government offices Both problems share the same root: no real-time, camera-first system that can take a photo and return a verified identity in seconds. ## Solution Three components, one platform: - **`apps/web`** — facial recognition identity verification. Captures a face photo via device camera, runs DeepFace analysis (age, gender, race, emotion), matches against a citizen database using ArcFace embeddings, and returns verified identity. Arabic UI with full RTL layout. - **`apps/vehicle`** — license plate recognition. Captures a vehicle photo, detects the car via YOLO object detection, crops and isolates the license plate, runs OCR to extract the plate number, and cross-references against registration data. Arabic UI. - **`backend`** — FastAPI server powering the facial recognition pipeline. Accepts base64 images, runs DeepFace analysis + face matching against a local database, returns structured identity results. ## Impact Both apps are **deployed on Vercel** and operational: - **Identity verification**: deployed at 3ain-sigma.vercel.app - **Vehicle recognition**: dep …