Reproduce “RLSplat: Resource-Linear Budget Control for Feed-forward 3D Gaussian Splatting” 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
Injae Kim, Minseok Joo, Jihwan Park, Sehyung Kim, Minseong Bae, Hyunwoo J. Kim
RLSplat reconstructs scenes from uncalibrated images by allocating a requested Gaussian budget to local regions before generation, then refining the Gaussians with local attention. Its decoder cost scales with the budget with fixed neighbor count. In zero-shot ACID testing, it reaches 25.11 dB PSNR versus 24.12 dB for F⁴Splat at 16K Gaussians; evaluation is at 256 × 256 in one cross-dataset setting.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.