Abstract:Deep neural networks (DNNs) are widely applied for nowadays 3D surface reconstruction tasks and such methods can be further divided into two categories, which respectively warp templates explicitly by moving vertices or represent 3D surfaces implicitly as signed or unsigned distance functions. Taking advantage of both advanced explicit learning process and powerful representation ability of implicit functions, we propose a novel 3D representation method, Neural Vector Fields (NVF). It not only adopts the explicit learning process to manipulate meshes directly, but also leverages the implicit representation of unsigned distance functions (UDFs) to break the barriers in resolution and topology. Specifically, our method first predicts the displacements from queries towards the surface and models the shapes as \textit{Vector Fields}. Rather than relying on network differentiation to obtain direction fields as most existing UDF-based methods, the produced vector fields encode the distance and direction fields both and mitigate the ambiguity at "ridge" points, such that the calculation of direction fields is straightforward and differentiation-free. The differentiation-free characteristic enables us to further learn a shape codebook via Vector Quantization, which encodes the cross-object priors, accelerates the training procedure, and boosts model generalization on cross-category reconstruction. The extensive experiments on surface reconstruction benchmarks indicate that our method outperforms those state-of-the-art methods in different evaluation scenarios including watertight vs non-watertight shapes, category-specific vs category-agnostic reconstruction, category-unseen reconstruction, and cross-domain reconstruction. Our code is released at <a class="link-external link-https" href="https://github.com/Wi-sc/NVF" rel="external noopener nofollow">this https URL</a>.

Unsigned Orthogonal Distance Fields: An Accurate Neural Implicit Representation for Diverse 3D Shapes

Learning Unsigned Distance Fields from Local Shape Functions for 3D Surface Reconstruction

NeuralUDF: Learning Unsigned Distance Fields for Multi-view Reconstruction of Surfaces with Arbitrary Topologies

HSDF: Hybrid Sign and Distance Field for Neural Representation of Surfaces with Arbitrary Topologies

NeUDF: Leaning Neural Unsigned Distance Fields with Volume Rendering

Neural Unsigned Distance Fields for Implicit Function Learning

Learning Density Regulated and Multi-View Consistent Unsigned Distance Fields.

SDFMAP: Neural Signed Distance Fields for Mapping and Positioning in Real-Time

DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling

Neural Vector Fields: Implicit Representation by Explicit Learning

Neural Surface Detection for Unsigned Distance Fields

A Deep Signed Directional Distance Function for Object Shape Representation

Details Enhancement in Unsigned Distance Field Learning for High-fidelity 3D Surface Reconstruction

Neural Vector Fields for Implicit Surface Representation and Inference

Probabilistic Directed Distance Fields for Ray-Based Shape Representations

Monge-Ampere Regularization for Learning Arbitrary Shapes from Point Clouds

CAP-UDF: Learning Unsigned Distance Functions Progressively from Raw Point Clouds with Consistency-Aware Field Optimization

Implicit Filtering for Learning Neural Signed Distance Functions from 3D Point Clouds

DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation

Quasi-Medial Distance Field (Q-MDF): A Robust Method for Approximating and Discretizing Neural Medial Axis