Inspiration
With the rise of biometric payments, facial recognition is becoming a common authentication method. However, this convenience comes with a growing threat: deepfakes, presentation attacks, and identity theft. Static facial authentication is no longer enough.
Our team — formed by members from four different nationalities — came together with one shared concern: how can we make biometric payments truly safe, globally and without friction?
That question led to BioTrust.
What it does
BioTrust is not a payment method.
It is an intelligent security layer that protects biometric payments by combining contextual risk analysis with biological proof of life (liveness detection).
Instead of forcing biometric checks on every transaction, BioTrust activates only when risk is detected.
Intelligent Risk Flow (Traffic Light System)
Low-risk scenario:
Known location + low transaction value
→ Payment approved instantly, without friction.High-risk scenario:
Unusual location or high transaction value
→ BioTrust is activated and requires biological validation.
How we built it
When a transaction is flagged as high-risk, BioTrust performs a Bio-Check using the device camera to verify that the user is a real, living human — not a photo, video, or deepfake.
Level 1 – Active Liveness Detection (MVP)
The system issues random challenges such as:
- Blink your eyes
- Turn your head
- Smile
We use MediaPipe Face Mesh and computer vision techniques to analyze facial geometry and motion in real time.
Level 2 – rPPG Heartbeat Detection (Experimental)
As a future enhancement, we explore remote photoplethysmography (rPPG), detecting subtle skin color changes caused by blood flow.
Our core message:
Screens don’t have blood. Deepfakes don’t have a pulse.
Tech stack
- Frontend: React / HTML + CSS
- Backend: Python (FastAPI)
- Computer Vision: MediaPipe, OpenCV
- Infrastructure: Local execution for low latency demos
Challenges we faced
- Lighting conditions affecting camera-based detection
- Avoiding unnecessary friction for legitimate users
- Balancing speed, accuracy, and user experience
We mitigated these risks by:
- Using active challenges instead of passive checks
- Designing a fallback demo system
- Prioritizing simplicity and robustness over over-complexity
What we learned
BioTrust showed us that the future of secure payments is not just digital — it’s biological.
By merging computer science with human physiology, we can create systems that scale globally while remaining resistant to synthetic identity attacks.
Log in or sign up for Devpost to join the conversation.