Figure1

Figure 1. The diagram of our proposed "Generative AI for RF Sensing" framework: After appropriate signal processing for different RF sensory data, GANs or diffusion models can start training and learning the diverse data distributions to generate high-quality synthetic data. The goal is to automatically and interactively (allowing users to provide prompts such as class labels and texts) generate, complete, enhance, edit or repair RF sensing data with fidelity and diversity.