GRNEJun 10, 2019

Differentiable Surface Splatting for Point-based Geometry Processing

arXiv:1906.04173v3380 citationsHas Code
Originality Highly original
AI Analysis

This work addresses geometry processing for point clouds, enabling inverse rendering without explicit connectivity, though it appears incremental as it builds on point-based methods.

The authors tackled the problem of high-fidelity differentiable rendering for point clouds by proposing Differentiable Surface Splatting (DSS), which accurately and robustly handles large-scale topological changes and small-scale detail modifications in applications like geometry synthesis and denoising, outperforming state-of-the-art techniques.

We propose Differentiable Surface Splatting (DSS), a high-fidelity differentiable renderer for point clouds. Gradients for point locations and normals are carefully designed to handle discontinuities of the rendering function. Regularization terms are introduced to ensure uniform distribution of the points on the underlying surface. We demonstrate applications of DSS to inverse rendering for geometry synthesis and denoising, where large scale topological changes, as well as small scale detail modifications, are accurately and robustly handled without requiring explicit connectivity, outperforming state-of-the-art techniques. The data and code are at https://github.com/yifita/DSS.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes