DBTB INT John Whaley r
summary
UnifyID demonstrates passive authentication by combining phone and IoT sensors to extract gait cadence from accelerometer streams and uniquely identify people with four seconds of walking data. The pipeline handles nonuniform sampling and heavy noise, uses signal processing to find step frequency, and matches that to a biometric template for authentication. The claim is reproducible: accelerometer-based gait yields high-precision identification after short windows when engineered around sensor idiosyncrasies.
AI-written summary · about
related talks
07mNot Quite Yourself Today: Behaviour-Based Continuous Authentication in IoT Environments
06mTouchID: User Authentication on Mobile Devices via Inertial-Touch Gesture Analysis
06mRF-Identity: Non-Intrusive Person Identification Based on Commodity RFID Devices
22mThe high cost of cheap biometrics
19mThumprint: Socially-Inclusive Local Group Authentication Through Shared Secret Knocks
25mBiometric proof of personhood verification in Worldcoin
20mBiometric Touch Sensing: Seamlessly Augmenting Each Touch with Continuous Authentication
27mEnsuring biometric recognition is secure and works well for all: A Case Study