Biometric Face Verification System

Biometric Face Verification System from scratch — here’s what it does.

Over the past few weeks I’ve been working on a complete identity verification system built with ASP.NET Core. No third-party service. No cloud dependency. Everything runs on your own server.

? The flow:

1️⃣ Register the user’s face once via a browser-based camera capture

2️⃣ Issue a verification link and send it to the user (SMS / email / in-app/browser)

3️⃣ The user completes a 5-step liveness challenge: Up → Down → Left → Right → Smile

4️⃣ Get a result back instantly with a similarity score

✅ Verified — similarity score: 0.84 to 0.97 (accuracy up to 97%)

❌ Failed — similarity score drops as low as 0.05, well below the acceptance threshold

⚙️ What’s under the hood:

→ 3 AI models working together seamlessly

→ A real-time face detector that works across angles and lighting conditions

→ A 468-point facial landmark model for precise head pose tracking

→ A deep embedding model trained on millions of identities for highly accurate face matching

→ Liveness detection to block photo and video spoofing attempts

→ WebSocket support for real-time frame processing

→ Token system with TTL and single-use enforcement

→ API key management with subscription plans (trial / monthly / yearly)

→ Full admin dashboard for managing client API keys and access control

→ Each client (site owner) can register users for facial recognition verification and easily track their verification status in real time

? Built as a service.

Any website owner can add face verification to their platform with a simple API key — no computer vision knowledge, no infrastructure to set up. The AI does the heavy lifting invisibly.

Under the hood it communicates through a REST API