MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction

By Zehao Yu, Songyou Peng, Michael Niemeyer, Torsten Sattler, Andreas Geiger. Neural Information Processing Systems, 2022

647 citations recorded by the source corpus.

In recent years, neural implicit surface reconstruction methods have become popular for multi-view 3D reconstruction. In contrast to traditional multi-view stereo methods, these approaches tend to produce smoother and more complete reconstructions due to the inductive smoothness bias of neural networks. State-of-the-art neural implicit methods allow for high-quality reconstructions of simple scenes from many input views. Yet, their performance drops significantly for larger and more complex scenes and scenes captured from sparse viewpoints. This is caused primarily by the inherent ambiguity in the RGB reconstruction loss that does not provide enough constraints, in particular in less-observed and textureless areas. Motivated by recent advances in the area of monocular geometry prediction, we systematically explore the utility these cues provide for improving neural implicit surface reconstruction. We demonstrate that depth and normal cues, predicted by general-purpose monocular estimators, significantly improve reconstruction quality and optimization time. Further, we analyse and investigate multiple design choices for representing neural implicit surfaces, ranging from monolithic MLP models over single-grid to multi-resolution grid representations. We observe that geometric monocular priors improve performance both for small-scale single-object as well as large-scale multi-object scenes, independent of the choice of representation.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/c6c518ccd441bae6fdc788a58b6d39db1f33f1a4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Computer Vision and Pattern Recognition

Citations in the corpus

UniSim: A Neural Closed-Loop Sensor Simulator (2023, methodology): “To improve the expressiveness and speed of NFFs, past works [12, 51, 74, 98] further combined learnable multi-resolution features grid {G l } Ll =1 with a neural network f .”
NoPe-NeRF: Optimising Neural Radiance Field with No Pose Prior (2022, methodology): “A number of techniques are proposed to improve NeRF’s performance with additional regularisation [13, 26, 54], depth priors [7,32,47,53], surface enhancements [27,43,49] or latent codes [40, 44, 52].”
Mip-Splatting: Alias-Free 3D Gaussian Splatting (2023, background): “3D radiance field reconstruction from multi-view observations is a well-known ill-posed problem as multiple distinctly different reconstructions can result in the same 2D projections [2, 55, 56].”
Gen3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control (2025, neutral): “Several works proposed to address these issues through regularization using geometric priors [10, 38, 43, 48, 54, 63, 70, 75], which, however, are sensitive to noise in the estimated depth or normals.”
GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Image (2024, methodology): “We conduct experiments on 4 scenes from ScanNet and employ evaluation metrics following [16,39,75].”
DNGaussian: Optimizing Sparse-View 3D Gaussian Radiance Fields with Global-Local Depth Normalization (2024, background): “The first group [11, 31] is to extract accurate but sparse depth values from reliable point clouds, and the second [14, 36, 40, 42, 55] distills depth knowledge from current powerful monocular depth estimators [29, 30].”
DUSt3R: Geometric 3D Vision Made Easy (2023, positive): “3D reconstruction from implicit models has undergone significant advancements, largely fueled by the integration of neural networks [61, 72, 81, 148, 173].”
NeRF-SLAM: Real-Time Dense Monocular SLAM with Neural Radiance Fields (2022, background): “While a variety of works have proposed to replace the density field by a signed distance functions or other representations[19], [20], [21], [22], we are most interested on the insights related to the use of alternative supervisory signals to increase the convergence speed of NeRFs, as well as the…”
SparseNeRF: Distilling Depth Ranking for Few-shot Novel View Synthesis (2023, methodology): “That is, we supervise a NeRF with relative depth instead of absolute depth [17, 65].”
Metric3D v2: A Versatile Monocular Geometric Foundation Model for Zero-Shot Metric Depth and Surface Normal Estimation (2024, neutral): “These unique attributes render both depth and surface normals indispensable in various computer vision applications, including 3D reconstruction [1], [2], [3], neural rendering (NeRF) [4], [5], [6], [7], autonomous driving [8], [9], [10], and robotics [11], [12], [13].”
MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View Images (2024, background): “Per-scene approaches mainly focus on designing effective regularization terms [8, 11, 26, 39, 43, 50] to better constrain the optimization process.”
ReconFusion: 3D Reconstruction with Diffusion Priors (2023, background): “Few-view NeRF Minimizing NeRF’s need for dense capture is crucial for democratizing 3D capture, and has motivated many works [10, 16, 22, 27, 35, 41, 47, 49, 50, 58, 62, 64, 65, 69].”
Neuralangelo: High-Fidelity Neural Surface Reconstruction (2023, background): “In contrast, our work Neuralangelo builds upon hash encodings [23] to recover surfaces but without the need for auxiliary inputs used in prior work [3, 5, 31, 44, 45].”
2D Gaussian Splatting for Geometrically Accurate Radiance Fields (2024, neutral): “Additional Key Words and Phrases: novel view synthesis, radiance fields, surface splatting, surface reconstruction”
Co-SLAM: Joint Coordinate and Sparse Parametric Encodings for Neural Real-Time SLAM (2023, methodology): “While these methods focus on novel view synthesis, others instead focus on surface reconstruction from RGB images, with implicit surface representations and differentiable renderers [8, 33, 39, 40].”