Supplementary Material for Shape As Points: A Differentiable Poisson Solver
–Neural Information Processing Systems
Next, we supply discussions on the easy initialization property of our Shape-As-Points representation in Section 3. Additional results and ablations for the optimization-based and We solve the PDEs using spectral methods [3]. Work done while at UC Berkeley. Please refer to Section 4.3 for a more in-depth discussion to motivate the use of the We optimize IGR for 15000 iterations on each object until convergence. For Point2Mesh, we follow the official implementation and use 6000 iterations for each object. We generate the initial mesh required by Point2Mesh following the description of the original paper.
Neural Information Processing Systems
Aug-15-2025, 01:03:03 GMT
- Country:
- Europe
- Switzerland > Zürich
- Zürich (0.04)
- Germany > Baden-Württemberg
- Tübingen Region > Tübingen (0.04)
- Switzerland > Zürich
- Europe
- Technology: