A Feature Adaptive Learning Method for High-Density sEMG-Based Gesture Recognition
summary
Presents a concrete method called FIDE: a stratified random forest ensemble that maintains per-electrode matrix sketches and a k-means electrode clustering to handle electrodes being added or lost, reconstructing missing channels via matrix sketch and updating trees for new channels. Reports improved accuracy over state of the art on three benchmark datasets (NinaPro, CSL-hdemg, CapgMyo) and a deployed multi-client/cloud system McS (smartphone + PC → cloud) that raises recognition accuracy during electrode increment and decrement.
AI-written summary · about
related talks
13mSparseEMG: Computational Design of Sparse EMG Layouts for Sensing Gestures
17mEMPress: Practical Hand Gesture Classification with Wrist-Mounted EMG and Pressure Sensing
20mAdvancing Hand Gesture Recognition with High Resolution Electrical Impedance Tomography
08mDesigning a Personalised Sensor Glove Using Deep-Learning
20mEarFieldSensing: A Novel In-Ear Electric Field Sensing to Enrich Wearable Gesture Input through ...
13mEnhanced EEG Data Classification with 3D SNN and On-Line Learning ESN
10mA large, crowdsourced evaluation of gesture generation systems on common data: The GENEA ...
26mBiosensor Machine Learning with Julia