Reproduce “RadPose: ISAC-based RF 3D Human Pose Estimation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Geo Jie Zhou, Kunzhe Song, Ruxin Lin, Huacheng Zeng
RadPose estimates a person’s 3D pose from a communication device’s own full-duplex OFDM transmissions, converting their reflections into radio point clouds and refining a skeleton across frames. On four unseen scenes in the authors’ dataset, it records 18.8 cm MPJPE versus 61.4 cm for mmDiff on the same dataset. This demonstrates held-out-scene performance in a prototype evaluation, not arbitrary-environment or commercial-device capability.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.