SynHAR
IMUDOI
Key facts
- Venue
- IEEE Access 2024
Summary
Amft-lab synthetic IMU augmentation method (IEEE Access 2024): biomechanical dynamics + human surface models → synthetic inertial data mixed into HAR train/pretrain. Precursor to Text2IMU (adds text-driven motion). Not a foundation encoder.
Source paper
SynHAR: Augmenting Human Activity Recognition With Synthetic Inertial Sensor Data Generated From Human Surface Models