Leveraging Dual-Observable Input for Fine-Grained Thumb Interaction Using Forearm EMG
summary
Prototype ThumbsUp uses seven dry-EMG channels plus accelerometer/gyroscope and Bluetooth to detect thumb gestures from the forearm, labeling EMG with phone-detected events via dual-observable input. A seven-participant study collects ~120 training points per user from lock‑screen and launcher interactions, recognizes left/right swipes, taps, long presses and user-defined gestures, and reports overall 82.9% accuracy (leave-session restriction drops to 67.6%, cross‑validation on training = 89.2%).
AI-written summary · about
chapters (8)
- 0:00Introduction to EMG interaction
- 2:12Dual input training approach
- 3:37Thumbs Up device demo
- 6:26User study evaluation
- 7:29Gesture frequency analysis
- 7:50Accuracy of thumb gesture classification
- 10:17Future applications and benefits
- 11:24Q&A session
related talks
17mEMPress: Practical Hand Gesture Classification with Wrist-Mounted EMG and Pressure Sensing
20mAdvancing Hand Gesture Recognition with High Resolution Electrical Impedance Tomography
16mInvestigating the feasibility of finger identification on capacitive touchscreens using deep ...
18mInfiniTouch: Finger-Aware Interaction on Fully Touch Sensitive Smartphones
06mWR-Hand: Wearable Armband can Track User's Hand
08mDesigning a Personalised Sensor Glove Using Deep-Learning
13mUniKey: Enabling Surface-Based Typing with Commodity Smartwatches via Cross-Modal Learning
28mStrike a Pose - Gesture recognition in JavaScript w/ ML & Arduino