Competitive Advantages
- Only accelerometer and gyroscope needed
- Easy software implementation
- Works in real-time without prior training
- Power efficient low sensor reading rate
Summary
USF inventors have developed a novel orientation invariant gait-matching algorithm that explicitly estimates the gait via the Kabsch approach using accelerometer and gyroscope sensors. The method was compared to the state-of-the-art using three different datasets with up to 745 subjects. The ability to recognize the owner while they do various activities with the phone placed in various places and distinguish the owner from other persons are compared. The algorithm significantly reduces error and is the best reported in the literature, thus enabling sensor-based gait authentication. This novel technology could be used for mobile biometrics, mobile authentication, activity recognition from inertial sensors, and health monitoring.
Operating Characteristic Curve With (red) and Without (blue) the Algorithm
Desired Partnerships
- License
- Sponsored Research
- Co-Development