Reproduce “Mesh-Learner: Texturing Mesh with Spherical Harmonics” 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
Yunfei Wan, Jianheng Liu, Chunran Zheng, Jiarong Lin, Fu Zhang
Mesh-Learner learns view-dependent appearance as spherical-harmonic coefficient textures on a supplied mesh, enabling training and rendering through a mesh-based rasterization pipeline. It transfers in-frustum textures to the GPU while keeping others in CPU memory. On Replica extrapolation, its average LPIPS is 0.047 versus 0.078 for M2-Mapping, though average PSNR is lower. This result is limited to that benchmark and protocol.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.