</> Code with Papers by Syntology

Datasets › ScanNet

ScanNet

Introduced by Angela Dai et al. in ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes1 Jan 2017 archive 2025-07-28

ScanNet is an instance-level indoor RGB-D dataset that includes both 2D and 3D data. It is a collection of labeled voxels rather than points or objects. Up to now, ScanNet v2, the newest version of ScanNet, has collected 1513 annotated scans with an approximate 90% surface coverage. In the semantic segmentation task, this dataset is marked in 20 classes of annotated 3D voxelized objects.

Source: A Review of Point Cloud Semantic Segmentation Image Source: http://www.scan-net.org/

HomepageBenchmarksPapers JSON

Benchmarks archive 2025-07-28

All 21 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Semantic Segmentation ScanNet DITR val mIoU 80.5 DINO in the Room: Leveraging 2D Foundation Models for 3D... VisualComputingInstitute/DITR 45 Compare
3D Object Detection ScanNetV2 DEST (based on V-DETR) (TTA) mAP@0.5 67.9 State Space Model Meets Transformer: A New Paradigm for... OpenSpaceAI/DEST3D 33 Compare
3D Instance Segmentation ScanNet(v2) Relation3D mAP @ 50 81.6 Relation3D : Enhancing Relation Modeling for Point Cloud... Howard-coder191/Relation3D 32 Compare
Semantic Segmentation ScanNetV2 CMX Mean IoU 61.3% CMX: Cross-Modal Fusion for RGB-X Semantic Segmentation... huaaaliu/rgbx_semantic_segmentation 12 Compare
3D Semantic Instance Segmentation ScanNetV2 3D-MPA mAP@0.50 61.1 3D-MPA: Multi Proposal Aggregation for 3D Semantic... francisengelmann/3D-MPA 5 Compare
Panoptic Segmentation ScanNetV2 OneFormer3D PQ 71.2 OneFormer3D: One Transformer for Unified Point Cloud Segmentation oneformer3d/oneformer3d 5 Compare
Panoptic Segmentation ScanNet OneFormer3D PQ 71.2 OneFormer3D: One Transformer for Unified Point Cloud Segmentation oneformer3d/oneformer3d 4 Compare
Depth Estimation ScanNetV2 Distill Any Depth absolute relative error 0.042 Distill Any Depth: Distillation Creates a Stronger... Westlake-AGI-Lab/Distill-Any-Depth 3 Compare
Scene Segmentation ScanNet 3DMV Average Accuracy 75.0% 3DMV: Joint 3D-Multi-View Prediction for 3D Semantic... angeladai/3DMV 3 Compare
Surface Normals Estimation ScanNetV2 Metric3Dv2 (g2, In-domain) % < 11.25 77.8 Metric3Dv2: A Versatile Monocular Geometric Foundation... yvanyin/metric3d 3 Compare
Depth Estimation ScanNet Atlas (plain) RMSE 0.165 Atlas: End-to-End 3D Scene Reconstruction from Posed Images magicleap/Atlas 2 Compare
Unsupervised 3D Semantic Segmentation ScanNetV2 Point-GCC+PointNet++ mIoU 18.3 Point-GCC: Universal Self-supervised 3D Scene... asterisci/point-gcc 2 Compare
2D Panoptic Segmentation ScanNetV2 EMSANet (2x ResNet-34 NBt1D, PanopticNDT version) PQ 58.22 PanopticNDT: Efficient and Robust Panoptic Mapping tui-nicr/emsanet +3 1 Compare
3D Instance Segmentation ScanNet MASC mAP 0.447 MASC: Multi-scale Affinity with Sparse Convolution for... art-programmer/MASC 1 Compare
3D Reconstruction ScanNet Atlas (finetuned) 3DIoU 89.4 Atlas: End-to-End 3D Scene Reconstruction from Posed Images magicleap/Atlas 1 Compare
3D Semantic Instance Segmentation ScanNetV1 SGPN mAP@0.25 35.1 SGPN: Similarity Group Proposal Network for 3D Point... laughtervv/SGPN 1 Compare
Continual Semantic Segmentation ScanNet LGKD mIoU 55.3 Label-Guided Knowledge Distillation for Continual... ze-yang/lgkd 1 Compare
Generalized Zero-Shot Learning ScanNet 3DGenZ HmIoU 12.5 Generative Zero-Shot Learning for Semantic Segmentation... valeoai/3DGenZ 1 Compare
Interactive 3D Instance Segmentation ScanNetV2 InterObject3D NoC@80 9.6 Interactive Object Segmentation in 3D Point Clouds theodorakontogianni/interobject3d 1 Compare
Interactive 3D Instance Segmentation -Trained on Scannet40 - Evaluated on Scannet40 ScanNetV2 InterObject3D NoC@80 8.9 Interactive Object Segmentation in 3D Point Clouds theodorakontogianni/interobject3d 1 Compare
Scene Recognition ScanNet SSMA Average Recall 54.28 Self-Supervised Model Adaptation for Multimodal Semantic... DeepSceneSeg/SSMA 1 Compare

Papers archive 2025-07-28

30 shown of 120 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 1,595. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentation 1 1 24 Mar 2025 not harvested
Sonata: Self-Supervised Learning of Reliable Point Representations 2 1 20 Mar 2025 ran 5 of 19 samples (14 unverified)
State Space Model Meets Transformer: A New Paradigm for 3D Object Detection 1 2 18 Mar 2025 ran 2 of 10 samples (8 unverified)
BFANet: Revisiting 3D Semantic Segmentation with Boundary Feature Analysis 1 1 16 Mar 2025 not harvested
Distill Any Depth: Distillation Creates a Stronger Monocular Depth Estimator 1 1 26 Feb 2025 not harvested
Beyond the Final Layer: Hierarchical Query Fusion Transformer with Agent-Interpolation Initialization for 3D Instance Segmentation 0 1 6 Feb 2025 not harvested
Relation3D : Enhancing Relation Modeling for Point Cloud Instance Segmentation 1 1 1 Jan 2025 not harvested
MSTA3D: Multi-scale Twin-attention for 3D Instance Segmentation 0 1 4 Nov 2024 not harvested
ARKit LabelMaker: A New Scale for Indoor 3D Scene Understanding 1 1 17 Oct 2024 ran 2 of 2 samples (0 unverified; 2 pointer-only for licence)
UniDet3D: Multi-dataset Indoor 3D Object Detection 1 1 6 Sep 2024 not harvested
MV-DETR: Multi-modality indoor object detection by Multi-View DEtecton TRansformers 0 1 13 Aug 2024 not harvested
Serialized Point Mamba: A Serialized Point Cloud Mamba Segmentation Model 0 2 17 Jul 2024 not harvested
Pamba: Enhancing Global Interaction in Point Clouds via State Space Model 0 1 25 Jun 2024 not harvested
KPConvX: Modernizing Kernel Point Convolution with Kernel Attention 1 1 21 May 2024 not harvested
Metric3Dv2: A Versatile Monocular Geometric Foundation Model for Zero-shot Metric Depth and Surface Normal Estimation 1 1 22 Mar 2024 ran 3 of 3 samples (0 unverified)
LSK3DNet: Towards Effective and Efficient 3D Perception with Large Sparse Kernels 1 1 22 Mar 2024 not harvested
OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation 1 1 21 Mar 2024 not harvested
AVS-Net: Point Sampling with Adaptive Voxel Size for 3D Scene Understanding 1 1 27 Feb 2024 not harvested
Scalable 3D Panoptic Segmentation As Superpoint Graph Clustering 1 2 12 Jan 2024 not harvested
ODIN: A Single Model for 2D and 3D Segmentation 1 2 4 Jan 2024 ran 3 of 8 samples (5 unverified)
SPGroup3D: Superpoint Grouping Network for Indoor 3D Object Detection 1 1 21 Dec 2023 not harvested
Spherical Mask: Coarse-to-Fine 3D Point Cloud Instance Segmentation with Spherical Representation 1 1 18 Dec 2023 ran 1 of 4 samples (3 unverified)
Point Transformer V3: Simpler, Faster, Stronger 3 1 15 Dec 2023 ran 7 of 11 samples (4 unverified)
OneFormer3D: One Transformer for Unified Point Cloud Segmentation 1 5 24 Nov 2023 not harvested
PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigm 1 1 12 Oct 2023 not harvested
PointHR: Exploring High-Resolution Architectures for 3D Point Cloud Segmentation 1 1 11 Oct 2023 not harvested
PanopticNDT: Efficient and Robust Panoptic Mapping 4 4 24 Sep 2023 ran 0 of 12 samples (12 unverified)
Mask-Attention-Free Transformer for 3D Instance Segmentation 1 1 4 Sep 2023 ran 4 of 10 samples (6 unverified; 10 pointer-only for licence)
Decoupled Local Aggregation for Point Cloud Learning 1 1 31 Aug 2023 not harvested
Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Training 1 1 18 Aug 2023 not harvested

The full list of 120 is in the JSON twin.

Dataset loaders archive 2025-07-28

Pointcept/Pointceptpytorch
www.scan-net.org/ScanNet/ScanNetpytorch

2 loaders as listed in the archive; links are outbound and not re-checked here.

Tasks archive 2025-07-28

Semantic Segmentation3D Object DetectionDepth Estimation3D ReconstructionPanoptic SegmentationScene SegmentationGeneralized Zero-Shot Learning3D Instance Segmentation2D Panoptic SegmentationScene RecognitionSurface Normals EstimationZero-shot 3D Point Cloud ClassificationContinual Semantic Segmentation3D Semantic Instance SegmentationInteractive 3D Instance SegmentationUnsupervised 3D Semantic SegmentationInteractive 3D Instance Segmentation -Trained on Scannet40 - Evaluated on Scannet40

License archive 2025-07-28

Custom

Modalities archive 2025-07-28

ImagesRGB-D

Languages archive 2025-07-28

No language tagged.

Variants archive 2025-07-28

  • ScanNet
  • ScanNetV1
  • ScanNetV2
  • ScanNet(v2)

4 variant names, as the archive lists them.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections