DiffCD: A Symmetric Differentiable Chamfer Distance for Neural Implicit Surface Fitting

• URL: http://arxiv.org/abs/2407.17058v1
• Date: Wed, 24 Jul 2024 07:36:33 GMT
• Title: DiffCD: A Symmetric Differentiable Chamfer Distance for Neural Implicit Surface Fitting
• Abstract summary: We show that state-of-the-art techniques work by minimizing an approximation of a one-sided Chamfer distance. As a consequence, existing methods can produce inaccurate reconstructions with spurious surfaces. We propose DiffCD, a novel loss function corresponding to the symmetric Chamfer distance.
• Score: 39.510392485356284
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Neural implicit surfaces can be used to recover accurate 3D geometry from imperfect point clouds. In this work, we show that state-of-the-art techniques work by minimizing an approximation of a one-sided Chamfer distance. This shape metric is not symmetric, as it only ensures that the point cloud is near the surface but not vice versa. As a consequence, existing methods can produce inaccurate reconstructions with spurious surfaces. Although one approach against spurious surfaces has been widely used in the literature, we theoretically and experimentally show that it is equivalent to regularizing the surface area, resulting in over-smoothing. As a more appealing alternative, we propose DiffCD, a novel loss function corresponding to the symmetric Chamfer distance. In contrast to previous work, DiffCD also assures that the surface is near the point cloud, which eliminates spurious surfaces without the need for additional regularization. We experimentally show that DiffCD reliably recovers a high degree of shape detail, substantially outperforming existing work across varying surface complexity and noise levels. Project code is available at https://github.com/linusnie/diffcd.

Related papers

• Learning Bijective Surface Parameterization for Inferring Signed Distance Functions from Sparse Point Clouds with Grid Deformation [50.26314343851213]
  Inferring signed distance functions (SDFs) from sparse point clouds remains a challenge in surface reconstruction. We present a novel approach that learns a dynamic deformation network to predict SDFs in an end-to-end manner. Experimental results on synthetic and real scanned datasets demonstrate that our method significantly outperforms the current state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-03-31T02:27:02Z)
• Thin-Shell-SfT: Fine-Grained Monocular Non-rigid 3D Surface Tracking with Neural Deformation Fields [66.1612475655465]
  3D reconstruction of deformable surfaces from RGB videos is a challenging problem. Existing methods use deformation models with statistical, neural, or physical priors. We propose ThinShell-SfT, a new method for non-rigid 3D tracking meshes.
  arXiv  Detail & Related papers  (2025-03-25T18:00:46Z)
• OffsetOPT: Explicit Surface Reconstruction without Normals [7.297352404640493]
  We propose OffsetOPT, a method that reconstructs explicit surfaces directly from 3D point clouds. We demonstrate its accuracy on popular benchmarks, including small-scale shapes and large-scale open surfaces.
  arXiv  Detail & Related papers  (2025-03-20T00:47:27Z)
• Geometry Field Splatting with Gaussian Surfels [23.412129038089326]
  We leverage the geometry field proposed in recent work for opaque surfaces, which can then be converted to volume densities. We adapt Gaussian kernels or surfels to the geometry field rather than the volume, enabling precise reconstruction of opaque solids. We demonstrate significant improvement in the quality of reconstructed 3D surfaces on widely-used datasets.
  arXiv  Detail & Related papers  (2024-11-26T03:07:05Z)
• Neural Octahedral Field: Octahedral prior for simultaneous smoothing and sharp edge regularization [9.167571374234166]
  We propose to guide surface reconstruction under a new variant of neural field, the octahedral field. By simultaneously fitting and smoothing the octahedral field alongside the implicit geometry, it behaves analogously to bilateral filtering. Our method outperforms various traditional and neural approaches across extensive experiments.
  arXiv  Detail & Related papers  (2024-08-01T06:02:59Z)
• Diffeomorphic Mesh Deformation via Efficient Optimal Transport for Cortical Surface Reconstruction [40.73187749820041]
  Mesh deformation plays a pivotal role in many 3D vision tasks including dynamic simulations, rendering, and reconstruction. A prevalent approach in current deep learning is the set-based approach which measures the discrepancy between two surfaces by comparing two randomly sampled point-clouds from the two meshes with Chamfer pseudo-distance. We propose a novel metric for learning mesh deformation, defined by sliced Wasserstein distance on meshes represented as probability measures that generalize the set-based approach.
  arXiv  Detail & Related papers  (2023-05-27T19:10:19Z)
• Minimal Neural Atlas: Parameterizing Complex Surfaces with Minimal Charts and Distortion [71.52576837870166]
  We present Minimal Neural Atlas, a novel atlas-based explicit neural surface representation. At its core is a fully learnable parametric domain, given by an implicit probabilistic occupancy field defined on an open square of the parametric space. Our reconstructions are more accurate in terms of the overall geometry, due to the separation of concerns on topology and geometry.
  arXiv  Detail & Related papers  (2022-07-29T16:55:06Z)
• Neural Reflectance for Shape Recovery with Shadow Handling [88.67603644930466]
  This paper aims at recovering the shape of a scene with unknown, non-Lambertian, and possibly spatially-varying surface materials. We propose a coordinate-based deep reflectance (multilayer perceptron) to parameterize both the unknown 3D shape and the unknown at every surface point. This network is able to leverage the observed photometric variance and shadows on the surface, and recover both surface shape and general non-Lambertian reflectance.
  arXiv  Detail & Related papers  (2022-03-24T07:57:20Z)
• MeshUDF: Fast and Differentiable Meshing of Unsigned Distance Field Networks [68.82901764109685]
  Recent work modelling 3D open surfaces train deep neural networks to approximate Unsigned Distance Fields (UDFs) We propose to directly mesh deep UDFs as open surfaces with an extension of marching cubes, by locally detecting surface crossings. Our method is order of magnitude faster than meshing a dense point cloud, and more accurate than inflating open surfaces.
  arXiv  Detail & Related papers  (2021-11-29T14:24:02Z)
• Differentiable Surface Triangulation [40.13834693745158]
  We present a differentiable surface triangulation that enables optimization for any per-vertex or per-face differentiable objective function over the space of underlying surface triangulations. Our method builds on the result that any 2D triangulation can be achieved by a suitably weighted Delaunay triangulation. We extend the algorithm to 3D by decomposing shapes into developable sets and differentiably meshing each set with suitable boundary constraints.
  arXiv  Detail & Related papers  (2021-09-22T12:42:43Z)
• DeepMesh: Differentiable Iso-Surface Extraction [53.77622255726208]
  We introduce a differentiable way to produce explicit surface mesh representations from Deep Implicit Fields. Our key insight is that by reasoning on how implicit field perturbations impact local surface geometry, one can ultimately differentiate the 3D location of surface samples. We exploit this to define DeepMesh -- end-to-end differentiable mesh representation that can vary its topology.
  arXiv  Detail & Related papers  (2021-06-20T20:12:41Z)
• MeshSDF: Differentiable Iso-Surface Extraction [45.769838982991736]
  We introduce a differentiable way to produce explicit surface mesh representations from Deep Signed Distance Functions. Our key insight is that by reasoning on how implicit field perturbations impact local surface geometry, one can ultimately differentiate the 3D location of surface samples. We exploit this to define MeshSDF, an end-to-end differentiable mesh representation which can vary its topology.
  arXiv  Detail & Related papers  (2020-06-06T23:44:05Z)
• SurfaceNet+: An End-to-end 3D Neural Network for Very Sparse Multi-view Stereopsis [52.35697180864202]
  Multi-view stereopsis (MVS) tries to recover the 3D model from 2D images. We investigate sparse-MVS with large baseline angles since the sparser sensation is more practical and more cost-efficient. We present SurfaceNet+, a volumetric method to handle the 'incompleteness' and the 'inaccuracy' problems induced by a very sparse MVS setup.
  arXiv  Detail & Related papers  (2020-05-26T13:13:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.