Zensei: Embedded, Multi-electrode Bioimpedance Sensing for Implicit, Ubiquitous User Recognition
summary
Custom multi-electrode bioimpedance hardware measures amplitude and phase across ~1 kHz–1.5 MHz and rotates up to six electrode grounds to capture 9000 measurements per sample. Prototypes embedded in a phone case, a grip pad and a chair collect data over a 22-day study with 46 subjects and train SVM classifiers. Hand grips reach about 95% accuracy while chair sensing is ~80–90% and false acceptance falls as low as 0.5% false acceptance for the phone.
AI-written summary · about
chapters (14)
- 0:00Introduction to Zensei
- 1:09How Zensei works
- 2:17Related work and inspiration
- 3:23Custom hardware design
- 4:46Sensor functionality
- 5:51Three form factors
- 6:55Demo of user recognition
- 7:31Advantages of sensing through clothing
- 8:24Long-term data collection
- 10:14Classification accuracy results
- 11:22Training data requirements
- 12:27Electrode arrangement importance
- 13:13Use cases and limitations
- 14:20Summary and future work
related talks
20mBiometric Touch Sensing: Seamlessly Augmenting Each Touch with Continuous Authentication
06mListen to Your Fingers: User Authentication based on Geometry Biometrics of Touch Gesture
06mTouchID: User Authentication on Mobile Devices via Inertial-Touch Gesture Analysis
19mEvaluating the Influence of Targets and Hand Postures on Touch-based Behavioural Biometrics
25mContinuous Biometrics
07mNot Quite Yourself Today: Behaviour-Based Continuous Authentication in IoT Environments
12mSkullConduct: Biometric User Identification on Eyewear Computers Using Bone Conduction Through ...
06mKEHKey: Kinetic Energy Harvester-based Authentication and Key Generation for Body Area Network