mTeeth: Identifying Brushing Teeth Surfaces UsingWrist-Worn Inertial Sensors
summary
Claims wrist accelerometer and gyroscope data can identify which of nine teeth surfaces receives brushing by detecting individual brushing strokes as valley-peak micro-events and classifying strokes with a layered Bayesian ensemble that models surface hierarchy and state transitions. Pipeline uses stroke-level features and correlation-based selection on a dataset with 10,230 labeled strokes from 114 episodes and achieves stroke-count error around 4.2% and counts within ±2, leave-one-subject F1 65%, leave-one-episode F1 80% and 10-fold F1 86%, with surface-duration MAE <7.5% (stroke anchor events).
AI-written summary · about
chapters (4)
related talks
17mEMPress: Practical Hand Gesture Classification with Wrist-Mounted EMG and Pressure Sensing
17mLearning to assess the quality of stroke rehabilitation exercises
09mAn Activity Recognition System for Taking Medicine Using In-The-Wild Data to Promote Medication ...
17mFeature Estimation for Punching Tool Wear at the Edge
20mAdvancing Hand Gesture Recognition with High Resolution Electrical Impedance Tomography
06mNeckSense: A Multi-Sensor Necklace for Detecting Eating Activities in Free-Living Conditions
17mPredicTouch: A System to Reduce Touchscreen Latency using Neural Networks and Inertial ...
06mVibroSense: Recognizing home activities by deep learning subtle vibrations on an interior ...