SwingNet: Ubiquitous Fine-Grained Swing Tracking Framework via Stochastic Neural Architecture ...
summary
A regression network discovered by stochastic neural architecture search fuses IMU and depth from a smart wristband (IMU) and smartphone (depth sensor), adds an adversarial domain discriminator and runs on phones. Prototypes track golf and tennis swings with a subject‑independent model that reports about 3.8 cm error in leave‑one‑subject tests and ~7.6 cm overall, under 10 cm claimed; mobile inference is energy‑and compute‑efficient and outdoor experiments show smooth predictions.
AI-written summary · about
related talks
24mProject GaitNet: Ushering in the ImageNet moment for human Gait kinematics
17mPredicTouch: A System to Reduce Touchscreen Latency using Neural Networks and Inertial ...
14mBack-Hand-Pose: 3D Hand Pose Estimation for a Wrist-worn Camera via Dorsum Deformation Network
11mBootstrapping User-Defined Body Tapping Recognition with Offline-Learned Probabilistic ...
08mSo Predictable! Continuous 3D Hand Trajectory Prediction in Virtual Reality
11mBaroPoser: Real-time Human Motion Tracking from IMUs and Barometers in Everyday Devices
08mDesigning a Personalised Sensor Glove Using Deep-Learning
20mInteracting with Soli: Exploring Fine-Grained Dynamic Gesture Recognition in the Radio-Frequency ...