Walk to Authenticate
Your walking pattern is your password. Enroll once, verify in 30 seconds.
How it works
Enroll
Walk normally for 60 seconds with your phone in your pocket. We learn your unique gait pattern.
Authenticate
Walk for 30 seconds. Your pattern is compared to your enrollment to verify your identity.
Result
You're either authenticated or not — no passwords needed.
About this research
This is a research project at Rijksuniversiteit Groningen exploring gait recognition as a biometric authentication method. Your smartphone's motion sensors (accelerometer and gyroscope) capture the unique patterns of your walk. Deep learning models analyze these patterns to create a personal gait signature that can verify your identity — a secure second factor for authentication.
Try different detection methods
All methods use the same sensor data but analyze it differently. Switch between them to find what works best for you.
Transformer
Uses a Transformer encoder to learn walking patterns from raw sensor data. Best overall accuracy.
LSTM
Uses an LSTM network to model the sequence of motion over time. Good for varied walking speeds.
FFT Centroids
Extracts frequency features from the motion signal and matches against stored centroids. Fastest inference.