SOTAVerified

3D Reconstruction

3D Reconstruction is the task of creating a 3D model or representation of an object or scene from 2D images or other data sources. The goal of 3D reconstruction is to create a virtual representation of an object or scene that can be used for a variety of purposes, such as visualization, animation, simulation, and analysis. It can be used in fields such as computer vision, robotics, and virtual reality.

Image: Gwak et al

Papers

Recently AddedMost HypedMost ActiveNeeds VerificationMost Verified

Showing 1–50 of 2326 papers

TitleStatusHype
TripoSR: Fast 3D Object Reconstruction from a Single ImageCode9
MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction PriorsCode7
Grounding Image Matching in 3D with MASt3RCode7
Instant Neural Graphics Primitives with a Multiresolution Hash EncodingCode6
MVSplat: Efficient 3D Gaussian Splatting from Sparse Multi-View ImagesCode5
PatchRefiner: Leveraging Synthetic Data for Real-Domain High-Resolution Monocular Metric Depth EstimationCode5
Structure-Aware Sparse-View X-ray 3D ReconstructionCode5
TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow ModelsCode5
Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse AttentionCode5
Neural Fields in Robotics: A SurveyCode5
InstantSplat: Sparse-view SfM-free Gaussian Splatting in SecondsCode5
DUSt3R: Geometric 3D Vision Made EasyCode5
SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB VideosCode5
ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View SynthesisCode5
Real3D-Portrait: One-shot Realistic 3D Talking Portrait SynthesisCode5
Prompting Depth Anything for 4K Resolution Accurate Metric Depth EstimationCode5
3D Reconstruction with Spatial MemoryCode5
RealFusion: 360° Reconstruction of Any Object from a Single ImageCode5
Infinite Photorealistic Worlds using Procedural GenerationCode5
Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward PassCode5
On Scaling Up 3D Gaussian Splatting TrainingCode4
Gaussian Splatting SLAMCode4
No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed ImagesCode4
LightGlue: Local Feature Matching at Light SpeedCode4
NeRDi: Single-View NeRF Synthesis with Language-Guided Diffusion as General Image PriorsCode4
Efficient LoFTR: Semi-Dense Local Feature Matching with Sparse-Like SpeedCode4
UniK3D: Universal Camera Monocular 3D EstimationCode4
Zero-1-to-3: Zero-shot One Image to 3D ObjectCode4
S^3Gaussian: Self-Supervised Street Gaussians for Autonomous DrivingCode4
DN-Splatter: Depth and Normal Priors for Gaussian Splatting and MeshingCode4
SpatialTrackerV2: 3D Point Tracking Made EasyCode4
Highly Accurate Dichotomous Image SegmentationCode4
GIM: Learning Generalizable Image Matcher From Internet VideosCode4
GRM: Large Gaussian Reconstruction Model for Efficient 3D Reconstruction and GenerationCode4
Real-time volumetric rendering of dynamic humansCode4
Cameras as Rays: Pose Estimation via Ray DiffusionCode4
GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single ImageCode4
Lotus: Diffusion-based Visual Foundation Model for High-quality Dense PredictionCode4
AGS-Mesh: Adaptive Gaussian Splatting and Meshing with Geometric Priors for Indoor Room Reconstruction Using SmartphonesCode4
One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape OptimizationCode3
PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian SplattingCode3
MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse ViewsCode3
MUSt3R: Multi-view Network for Stereo 3D ReconstructionCode3
Anything-3D: Towards Single-view Anything Reconstruction in the WildCode3
MVSFormer++: Revealing the Devil in Transformer's Details for Multi-View StereoCode3
PE3R: Perception-Efficient 3D ReconstructionCode3
LiDAR4D: Dynamic Neural Fields for Novel Space-time View LiDAR SynthesisCode3
DOGS: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian ConsensusCode3
EscherNet: A Generative Model for Scalable View SynthesisCode3
Large Spatial Model: End-to-end Unposed Images to Semantic 3DCode3
Show:102550
← PrevPage 1 of 47Next →
All datasetsDTUShapeNetData3D−R2N2Scan2CAD3DPeopleAria Digital Twin DatasetAria Synthetic EnvironmentsScanNet

Benchmark Results

#ModelMetricClaimedVerifiedStatus
13D-R2N2Overall0.63—Unverified
2GipumaOverall0.58—Unverified
3COLMAPOverall0.53—Unverified
4MVSNetOverall0.46—Unverified
5Vis-MVSNetOverall0.37—Unverified
6AA-RMVSNetOverall0.36—Unverified
7Cas-MVSNetOverall0.36—Unverified
8EPP-MVSNetOverall0.36—Unverified
9PatchmatchNetOverall0.35—Unverified
10CVP-MVSNetOverall0.35—Unverified
#ModelMetricClaimedVerifiedStatus
1MD-GONIoU92.8—Unverified
2POCOIoU92.6—Unverified
3FS-SDFIoU91.2—Unverified
4DP-ConvONetIoU89.5—Unverified
5ConvONetIoU88.4—Unverified
6ONetIoU76.1—Unverified
7EVolTIoU73.8—Unverified
8ZubicLioIoU65.43—Unverified
#ModelMetricClaimedVerifiedStatus
1AttSets3DIoU0.64—Unverified
2PSGN3DIoU0.64—Unverified
3OGN3DIoU0.6—Unverified
43D-R2N23DIoU0.56—Unverified
#ModelMetricClaimedVerifiedStatus
1Scan2CADAverage Accuracy31.68—Unverified
23DMatchAverage Accuracy10.29—Unverified
#ModelMetricClaimedVerifiedStatus
1SVCPChamfer10—Unverified
#ModelMetricClaimedVerifiedStatus
1EVLAccuracy18.2—Unverified
#ModelMetricClaimedVerifiedStatus
1EVLAccuracy5.7—Unverified
#ModelMetricClaimedVerifiedStatus
1Atlas (finetuned)3DIoU89.4—Unverified