Noisy Labels Make Sense: Data-Driven Smartphone Inertial Tracking without Tedious Annotations.
Yuefan TongJiankun WangZenghua ZhaoJiayang CuiBin WuPublished in: WoWMoM (2025)
Keyphrases
- data driven
- kalman filter
- real time
- particle filtering
- error prone
- data sets
- ground truth labels
- inertial sensors
- motion tracking
- dynamic model
- motion segmentation
- visual tracking
- motion model
- pairwise
- mean shift
- object tracking
- particle filter
- mobile devices
- metadata
- missing data
- image classification
- optical flow
- video sequences
- mobile applications
- training data
- position and orientation
- computer vision
- learning algorithm