Reproduce “GTSM: Gaussian Splatting with Thermal State-Space Dynamics for Spatio-Temporal Thermal Field Reconstruction and Prediction” 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
Yunpeng Zhao, Haoxiang Zhang, Chao Liu
GTSM reconstructs and forecasts thermal fields from infrared views over time. It attaches continuously evolving thermal states to Gaussian primitives, combining a shared scene-level driver with local responses, without prescribing a heat-evolution law. On RHD, its mean extrapolation PSNR is 40.98 dB, versus 38.11 dB for a direct MLP and 29.49 dB for ETGS. Real sequences have approximately static geometry; deforming-target evidence is synthetic.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.