Geometry Reconstruction and Denoising

The field of geometry reconstruction and denoising is moving towards more efficient and accurate methods for recovering high-quality surfaces from point clouds and other data. Recent developments have focused on incorporating prior knowledge and geometric priors into the reconstruction process, allowing for more robust and detailed results. Notable papers in this area include:

  • Self-Supervised Implicit Attention Priors for Point Cloud Reconstruction, which introduces a novel approach to distilling shape-specific priors directly from the input point cloud.
  • A Finite Difference Approximation of Second Order Regularization of Neural-SDFs, which proposes a finite-difference framework for curvature regularization in neural signed distance field learning, offering an efficient and scalable alternative for curvature-aware SDF learning.

Sources

Geometry Denoising with Preferred Normal Vectors

Self-Supervised Implicit Attention Priors for Point Cloud Reconstruction

Implicit reconstruction from point cloud: an adaptive level-set-based semi-Lagrangian method

Constructive quasi-uniform sequences over triangles

Accurate and Efficient Surface Reconstruction from Point Clouds via Geometry-Aware Local Adaptation

A Finite Difference Approximation of Second Order Regularization of Neural-SDFs

Built with on top of