Reproduce “Neural Implicit Energy Optimization for LiDAR Mapping” 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
Hrishikesh Viswanath, Md Ashiqur Rahman, Chi Lin, Damon Conover, Aniket Bera
NIEO adds a surface-curvature penalty to neural signed-distance field training, using shape-operator energy to favor smooth LiDAR maps. It retains data-fit and Eikonal constraints rather than smoothing extracted meshes. On Newer College, it raises F-score from 93.70% to 94.41% and improves accuracy, but worsens completeness from 10.00 cm to 11.18 cm. This adds training computation, with no added inference-time cost.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.