M4X: Enhancing Cross-View Generalizability in RF-Based Human Activity Recognition by Exploiting Synthetic Data in Metric Learning.
Mengjing LiuZongxing XieFan YePublished in: CHASE (2024)
Keyphrases
- synthetic data
- metric learning
- human activity recognition
- cross view
- person re identification
- activity recognition
- human activities
- distance metric
- distance function
- similarity search
- semi supervised
- pairwise
- multi task
- real world
- graph embedding
- dimensionality reduction
- learning tasks
- data sets
- feature space
- active learning
- semi supervised learning
- machine learning
- feature extraction