USAD: End-to-End Human Activity Recognition via Diffusion Model with Spatiotemporal Attention

Open in new window