TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Datasets/ScanNet++

ScanNet++

ScanNet++: A High-Fidelity Dataset of 3D Indoor Scenes

ScanNet++ is a large scale dataset with 450+ 3D indoor scenes containing sub-millimeter resolution laser scans, registered 33-megapixel DSLR images, and commodity RGB-D streams from iPhone. The 3D reconstructions are annotated with long-tail and label-ambiguous semantics to benchmark semantic understanding methods, while the coupled DSLR and iPhone captures enable benchmarking of novel view synthesis methods in high-quality and commodity settings.

Benchmarks

10-shot image generation/Top-1 IoU10-shot image generation/Top-3 IoU16k/mAP@0.2516k/mAP@0.52D Classification/mAP@0.252D Classification/mAP@0.52D Object Detection/mAP@0.252D Object Detection/mAP@0.53D/mAP@0.253D/mAP@0.53D Instance Segmentation/mAP3D Object Detection/mAP@0.253D Object Detection/mAP@0.53D Semantic Segmentation/Top-1 IoU3D Semantic Segmentation/Top-3 IoUInstance Segmentation/mAPNovel View Synthesis/PSNRNovel View Synthesis/SSIMNovel View Synthesis/LPIPSObject Detection/mAP@0.25Object Detection/mAP@0.5Semantic Segmentation/Top-1 IoUSemantic Segmentation/Top-3 IoU

Related Benchmarks

ScanNet++ (trained on 3DMatch)/3D Point Cloud Interpolation/Recall ( correspondence RMSE below 0.2)ScanNet++ (trained on 3DMatch)/Point Cloud Registration/Recall ( correspondence RMSE below 0.2)

Statistics

Papers
25
Benchmarks
23

Links

Homepage

Tasks

10-shot image generation16k2D Classification2D Object Detection3D3D Instance Segmentation3D Object Detection3D Semantic SegmentationInstance SegmentationNovel View SynthesisObject DetectionPoint Cloud RegistrationSemantic Segmentation