Meshlet Priors for 3D Mesh Reconstruction
This addresses the challenge of 3D mesh reconstruction for applications like computer graphics and robotics by providing a more generalizable and pose-invariant method compared to existing class-specific deep-learning approaches.
The paper tackles the problem of reconstructing 3D meshes from sparse, noisy point clouds by introducing meshlets, small patches that learn local shape priors, enabling reconstruction of objects in any pose and from unseen classes with improved robustness to noise and sparsity.
Estimating a mesh from an unordered set of sparse, noisy 3D points is a challenging problem that requires carefully selected priors. Existing hand-crafted priors, such as smoothness regularizers, impose an undesirable trade-off between attenuating noise and preserving local detail. Recent deep-learning approaches produce impressive results by learning priors directly from the data. However, the priors are learned at the object level, which makes these algorithms class-specific and even sensitive to the pose of the object. We introduce meshlets, small patches of mesh that we use to learn local shape priors. Meshlets act as a dictionary of local features and thus allow to use learned priors to reconstruct object meshes in any pose and from unseen classes, even when the noise is large and the samples sparse.