Signed Distance Functions

Also known as SDF

A signed distance function assigns each point its distance to the nearest surface, with the sign indicating which side of the surface it lies on; the zero level set represents the surface itself. Researchers learn these fields from images, depth maps, and point clouds using neural implicit networks, spatial grids, and hybrid encodings, often adding geometric priors or uncertainty estimates. In robotics, distance values and gradients support collision checking, motion planning, and safety control, while in vision they support surface reconstruction and shape analysis. A central challenge is preserving accurate distances, signs, and thin or open geometry while keeping representations efficient and reliable with incomplete data, especially near medial axes and boundaries.

Latest papers

38 papers3 papers in the last four weeks, down 25% on the four weeks before. 0.0% of all new papers.

Fus3D: Decoding Consolidated 3D Geometry from Feed-forward Geometry Transformer Latents
Mar 26, 2026Laura Fink, Linus Franke, George Kopanas +2Feed-Forward 3D Reconstruction Multi-View 3D Reconstruction
Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany · Inria, Université Côte d’Azur, France · Google DeepMind, UK
cuRoboV2: Dynamics-Aware Motion Generation with Depth-Fused Distance Fields for High-DoF Robots
Mar 5, 2026Balakumar Sundaralingam, Adithyavairavan Murali, Stan BirchfieldRobot Motion Generation Kinodynamic Motion Planning
Leveraging Geometric Prior Uncertainty and Complementary Constraints for High-Fidelity Neural Indoor Surface Reconstruction
Feb 27, 2026Qiyu Feng, Jiwei Shan, Shing Shin Cheng +1UQ for 3D Reconstruction 3D Reconstruction
Department of Automation, Shanghai Jiao Tong University, Shanghai 200240, China. · Department of Mechanical and Automation Engineering and T Stone Robotics Institute, The Chinese University of Hong Kong, Hong Kong.
SQ-CBF: Signed Distance Functions for Reliable Superquadric-Based Safety Filtering
Feb 11, 2026Haocheng Zhao, Lukas Brunke, Oliver Lagerquist +2Safety Filtering Control Barrier Functions
Learning Systems and Robotics Lab, Technical University of Munich, 80333 Munich, Germany, and also with Simon Fraser University, Burnaby, BC V5A 1S6, Canada
TetraSDF: Analytic Isosurface Extraction with Multi-resolution Tetrahedral Grid
Nov 20, 2025Seonghun Oh, Youngjung Uh, Jin-Hwa Kim3D Mesh Generation Neural Representation Geometry
Yonsei University, Republic of Korea · NAVER AI Lab, Republic of Korea · Seoul National University, Republic of Korea
OREN: Octree Residual Network for Real-Time Euclidean Signed Distance Mapping
Oct 21, 2025Zhirui Dai, Qihao Qian, Tianxing Fan +13D Mapping Implicit Neural Representations
Department of Electrical and Computer Engineering, University of California San Diego, La Jolla, CA 92093, USA
VesselSDF: Distance Field Priors for Vascular Network Reconstruction
Jun 19, 2025Salvatore Esposito, Daniel Rebain, Arno Onken +2Medical Image Segmentation Signed Distance Functions
School of Informatics, University of Edinburgh, Edinburgh, United Kingdom · Department of Computer Science, University of British Columbia, Vancouver, Canada
GSurf: Learning Signed Distance Fields from Splatting Opaque Gaussians for High-quality 3D Reconstruction
Nov 24, 2024Baixin Xu, Jiangbei Hu, Jiaze Li +13D Gaussian Splatting Signed Distance Functions
College of Computing and Data Science, Nanyang Technological University · School of Software, Dalian University of Technology

Mar 26, 2026J/K move · Enter open · S save

Fus3D: Decoding Consolidated 3D Geometry from Feed-forward Geometry Transformer Latents

Laura Fink, Linus Franke, George Kopanas, Marc Stamminger+1

Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany · Inria, Université Côte d’Azur, France · Google DeepMind, UK

Feed-Forward 3D Reconstruction Multi-View 3D Reconstruction Sparse-View 3D Reconstruction Signed Distance Functions 3D ViTs

Open paper AskPDF arXiv

We propose a feed-forward method for dense Signed Distance Field (SDF) regression from unstructured image collections in less than three seconds, without camera calibration or post-hoc fusion. Our key insight is that the intermediate feature space of pretrained multi-view feed-forward geometry transformers already encodes a powerful joint world representation; yet, existing pipelines discard it, routing features through per-view prediction heads before assembling 3D geometry post-hoc, which discards valuable completeness information and accumulates inaccuracies. We instead perform 3D extraction directly from geometry transformer features via learned volumetric extraction: voxelized canonical embeddings that progressively absorb multi-view geometry information through interleaved cross- and self-attention into a structured volumetric latent grid. A simple convolutional decoder then maps this grid to a dense SDF. We additionally propose a scalable, validity-aware supervision scheme directly using SDFs derived from depth maps or 3D assets, tackling practical issues like non-watertight meshes. Our approach yields complete and well-defined distance values across sparse- and dense-view settings and demonstrates geometrically plausible completions. Code and further material can be found at https://lorafib.github.io/fus3d.