Reflector: Distance-Independent, Private Pointing on a Reflective Screen
summary
Uses a front camera at the screen edge to map face features into an onscreen reflection cursor by estimating scale and translation and updating scale from RGB or depth-derived distance. Requires a one-time calibration of about 1.5 minutes and runs on commodity glossy screens. Works reliably up to 2.5 m with lower far-distance error than an eye tracker, and yields 20× lower shoulder-surfing success than touch for 4-digit PIN entry.
AI-written summary · about
chapters (8)
- 0:00Introduction to Reflector
- 2:06Limitations of current input
- 4:22System overview
- 5:38Calibration and tracking
- 7:38Real-time scaling process
- 8:39Pointing performance study
- 12:22Authentication security
- 13:56Discussion and Q&A
related talks
12mPricoEye: The Eye of Primary Colors for Fast and Convenient 3D Reconstruction of Fine-grained Pal...
34mHow Eyetracking can be Beginning of the End of Privacy
45mYour passport to the 7th continent
14mRed Bleed: A Pragmatic Near-Infrared Presentation Attack on Facial Biometric...
06mIriTrack: Face Presentation Attack Detection Using Iris Tracking
16mYour FACE in a SPACE
06mTouchID: User Authentication on Mobile Devices via Inertial-Touch Gesture Analysis
24mMy Google Glass Sees Your Passwords!