Reproduce “T$^2$-Splat: Adaptive Topology Mesh Splatting with Texture Residuals” 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
Zhihao Tang, Youjia Zhang, Mingbo Zhao, Wei Yang
T²-Splat adapts mesh topology during optimization, then adds per-face residual textures to preserve appearance detail without further changing geometry. On Mip-NeRF 360, it reports PSNR 25.49 dB, LPIPS 0.234, and SSIM 0.779, versus MeshSplatting’s 24.78 dB, 0.310, and 0.728. On Garden, reducing the face budget to 10% lowers PSNR to 23.93 dB, showing a quality trade-off.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.