Synchronized raw radar and three-dimensional human pose data for prehabilitation exercises

We present a synchronized dataset of millimeter-wave radar measurements and camera-derived 3D human pose annotations acquired during prehabilitation exercises. Thirteen volunteers performed nineteen standing activities targeting strength, flexibility, balance, and core stability following a standardized recording protocol. Radar data were collected with a four-dimensional in-phase and quadrature millimeter-wave radar configured with 10 transmit antennas and 20 receivers and operated in stepped-frequency continuous-wave mode over 62–66.5 GHz with 100 frequency tones, yielding a frame rate of approximately 13 Hz. In parallel, an RGB-D camera recorded video at 30 fps, and 33-joint 3D skeletal landmarks were extracted offline using MediaPipe BlazePose Full and remapped to a COCO-style 17-joint representation. The radar and pose streams were temporally aligned by nearest-neighbor timestamp matching. The dataset includes subject-level files containing IQ radar data, camera timestamps with alignment indices, and 3D joint coordinates in a local hip-centered coordinate system. These data may support research on radar-based human pose estimation and multimodal sensing for rehabilitation-related applications.