IM360: Large-scale Indoor Mapping with 360 Cameras
Not on the map. Why?
Abstract (arXiv)
We present a novel 3D mapping pipeline for large-scale indoor environments. To address the significant challenges in large-scale indoor scenes, such as prevalent occlusions and textureless regions, we propose IM360, a novel approach that leverages the wide field of view of omnidirectional images and integrates the spherical camera model into the Structure-from-Motion (SfM) pipeline. Our SfM utilizes dense matching features specifically designed for 360 images, demonstrating superior capability in image registration. Furthermore, with the aid of mesh-based neural rendering techniques, we introduce a texture optimization method that refines texture maps and accurately captures view-dependent properties by combining diffuse and specular components. We evaluate our pipeline on large-scale indoor scenes, demonstrating its effectiveness in real-world scenarios. In practice, IM360 demonstrates superior performance, achieving a 3.5 PSNR increase in textured mesh reconstruction. We attain state-of-the-art performance in terms of camera localization and registration on Matterport3D and Stanford2D3D.
Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 49 of its 78 references that are in the map. First posted on arXiv 18 Feb 2025.
Builds on
49 of 78 references.
References · 49
- 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
- 20233D Gaussian Splatting for Real-Time Radiance Field Renderingcited by 6,226
- 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
- 2021Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fieldscited by 2,032
- 2021NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstructioncited by 1,712
- 2019SuperGlue: Learning Feature Matching with Graph Neural Networkscited by 1,443
- 2017Matterport3D: Learning from RGB-D Data in Indoor Environmentscited by 1,376
- 2017SuperPoint: Self-Supervised Interest Point Detection and Descriptioncited by 1,320
- 2021LoFTR: Detector-Free Local Feature Matching with Transformerscited by 984
- 2020Accelerating 3D Deep Learning with PyTorch3Dcited by 784
- 2020Implicit Geometric Regularization for Learning Shapescited by 710
- 2023Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fieldscited by 671
37 more
- 2023SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Renderingcited by 616
- 2021Habitat-Matterport 3D Dataset (HM3D): 1000 Large-scale 3D Environments for Embodied AIcited by 592
- 2018From Coarse to Fine: Robust Hierarchical Localization at Large Scalecited by 559
- 2022MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstructioncited by 470
- 2021Baking Neural Radiance Fields for Real-Time View Synthesiscited by 468
- 2021ARKitScenes: A Diverse Real-World Dataset For 3D Indoor Scene Understanding Using Mobile RGB-D Datacited by 454
- 2020Modular Primitives for High-Performance Differentiable Renderingcited by 441
- 2017Joint 2D-3D-Semantic Data for Indoor Scene Understandingcited by 406
- 2021Extracting Triangular 3D Models, Materials, and Lighting From Imagescited by 366
- 2021Omnidata: A Scalable Pipeline for Making Multi-Task Mid-Level Vision Datasets from 3D Scanscited by 336
- 2020DISK: Learning local features with policy gradientcited by 312
- 2022MobileNeRF: Exploiting the Polygon Rasterization Pipeline for Efficient Neural Field Rendering on Mobile Architecturescited by 311
- 2023RoMa: Robust Dense Feature Matchingcited by 302
- 2023SparseNeRF: Distilling Depth Ranking for Few-shot Novel View Synthesiscited by 247
- 2024Global Structure-from-Motion Revisitedcited by 218
- 2022DKM: Dense Kernelized Feature Matching for Geometry Estimationcited by 177
- 2022ASpanFormer: Detector-Free Image Matching with Adaptive Span Transformercited by 171
- 2019GLU-Net: Global-Local Universal Network for Dense Flow and Correspondencescited by 139
- 2016Image-based localization using LSTMs for structured feature correlationcited by 136
- 2023SparseGS: Sparse View Synthesis using 3D Gaussian Splattingcited by 134
- 2021Learning Accurate Dense Correspondences and When to Trust Themcited by 113
- 2019Tangent Images for Mitigating Spherical Distortioncited by 79
- 2023Detector-Free Structure from Motioncited by 71
- 2018DGC-Net: Dense Geometric Correspondence Networkcited by 67
- 2023Urban Radiance Field Representation with Deformable Neural Mesh Primitivescited by 55
- 2024VCR-GauS: View Consistent Depth-Normal Regularizer for Gaussian Surface Reconstructioncited by 50
- 2023Real-Time Neural Rasterization for Large Scenescited by 37
- 2021Large-scale Localization Datasets in Crowded Indoor Spacescited by 29
- 2023Balanced Spherical Grid for Egocentric View Synthesiscited by 29
- 2021Differentiable Stereopsis: Meshes from multiple views using differentiable renderingcited by 26
- 2023DebSDF: Delving into the Details and Bias of Neural Indoor Scene Reconstructioncited by 17
- 2024Feature Splatting for Better Novel View Synthesis with Low Overlapcited by 9
- 20233D reconstruction from spherical images: A review of techniques, applications, and prospectscited by 7
- 2023TMO: Textured Mesh Acquisition of Objects with a Mobile Device by using Differentiable Renderingcited by 7
- 2024OmniSDF: Scene Reconstruction using Omnidirectional Signed Distance Functions and Adaptive Binoctreescited by 6
- 2024Mode-GS: Monocular Depth Guided Anchored 3D Gaussian Splatting for Robust Ground-View Scene Renderingcited by 5
- 2021Robust 360-8PA: Redesigning The Normalized 8-point Algorithm for 360-FoV Imagescited by 3
Built on by
Citing papers per year: 2026: 3.
3 papers.