SkullConduct: Biometric User Identification on Eyewear Computers Using Bone Conduction Through ...
summary
SkullConduct uses a bone‑conduction speaker on eyewear and the device microphone to play and record a white‑noise probe, extracts MFCC features and classifies with a lightweight 1‑NN, reporting identification at 97.0% accuracy and authentication with an equal error rate of 6.9%. A 23‑second probe was used in a controlled lab with 10 participants (one second suffices for recognition), and limitations include background noise, placement variance and unknown scalability to many users.
AI-written summary · about
chapters (6)
- 0:00Introduction to SkullConduct
- 1:51How SkullConduct works
- 3:01System implementation details
- 4:26Study setup and results
- 7:39Take home and future work
- 8:01Q&A session
related talks
06mEarDynamic: An Ear Canal Deformation Based Continuous User Authentication Using In-Ear Wearables
16mEchoLLM: LLM-Augmented Acoustic Eavesdropping Attack on Bone Conduction...
33mFor Smarter Authentication, We Might Need to Use ...
06mVoice In Ear: Spoofing-Resistant and Passphrase-Independent Body Sound Authentication
20mBiometric Touch Sensing: Seamlessly Augmenting Each Touch with Continuous Authentication
18mZensei: Embedded, Multi-electrode Bioimpedance Sensing for Implicit, Ubiquitous User Recognition
06mListen to Your Fingers: User Authentication based on Geometry Biometrics of Touch Gesture
25mContinuous Biometrics