Reproduce “StickFigure: Enabling Pose-Only IoT Cameras for Privacy-Preserving Edge Vision” 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

Aveline Xiaoyuan Han, Chenshu Wu

StickFigure performs pose estimation on an MCU camera and exports joint coordinates instead of monitoring-time RGB. It keeps early high-resolution feature channels narrow, adds capacity after downsampling, and refines elbow and wrist estimates. On NTU RGB+D, it reports 78.40% [email protected] and 74.99±1.06% HAR Top-1, outperforming its XiNet-based baseline; this is not a formal privacy guarantee.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.