The major obstacle for learning-based RF sensing is to obtain a high-quality large-scale annotated dataset. However, unlike visual datasets that can be easily annotated by human workers, RF signal is non-intuitive and non-interpretable, which causes the annotation of RF signals time-consuming and laborious. To resolve the rapacious appetite of anno...
|