Reproduce “NeuralPCI: Spatio-temporal Neural Field for 3D Point Cloud Multi-frame Non-linear Interpolation” 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
Zehan Zheng, Danni Wu, Ruisi Lu, Fan Lu, Guang Chen, Changjun Jiang
NeuralPCI uses neighboring LiDAR point clouds to fit a continuous motion representation across space and time, then queries it for moments between captures. On NL-Drive, a motion-focused stress test, its average Chamfer Distance (CD) is 0.80 across three equally spaced interpolated frames, versus 1.06 for PointINet. This supports lower CD on this benchmark, not broad real-world generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.