List your company
ResearchFormerly WizWand

3D Object Detection on ScanNet 60

32mAP@0.25Zoo3D1
-1.0727.51416.124.686Mar 23, 2026
Updated 7mo ago

Evaluation results

MethodLinks
Zoo3D1Input modality=Point cloud-based, Pose-free=N/A, Zero-shot=true, Use 3D bounding boxes during training=false2026.033220.8
Group3DInput modality=Multi-view image-based, Pose-free=false, Zero-shot=false, Use 3D bounding boxes during training=false2026.0329.113.9
Zoo3D0Input modality=Point cloud-based, Pose-free=N/A, Zero-shot=true, Use 3D bounding boxes during training=false2026.0327.118.7
Zoo3D1Input modality=Multi-view image-based, Pose-free=false, Zero-shot=false, Use 3D bounding boxes during training=false2026.0323.910.8
Group3DInput modality=Multi-view image-based, Pose-free=true, Zero-shot=true, Use 3D bounding boxes during training=false2026.0322.38.5
Zoo3D0Input modality=Multi-view image-based, Pose-free=false, Zero-shot=false, Use 3D bounding boxes during training=false2026.032210.4
OV-Uni3DETRInput modality=Point cloud-based, Pose-free=N/A, Zero-shot=false, Use 3D bounding boxes during training=true2026.0319.4—
Zoo3D1Input modality=Multi-view image-based, Pose-free=true, Zero-shot=true, Use 3D bounding boxes during training=false2026.0315.35.6
Zoo3D0Input modality=Multi-view image-based, Pose-free=true, Zero-shot=true, Use 3D bounding boxes during training=false2026.0313.34.1
OV-Uni3DETRInput modality=Multi-view image-based, Pose-free=false, Zero-shot=false, Use 3D bounding boxes during training=true2026.0311.2—
INHAInput modality=Point cloud-based, Pose-free=N/A, Zero-shot=false, Use 3D bounding boxes during training=true2026.0310.7—
CoDaInput modality=Point cloud-based, Pose-free=N/A, Zero-shot=false, Use 3D bounding boxes during training=true2026.039—
3D-CLIPInput modality=Point cloud-based, Pose-free=N/A, Zero-shot=false, Use 3D bounding boxes during training=true2026.034—
Det-PointCLIPv2Input modality=Point cloud-based, Pose-free=N/A, Zero-shot=false, Use 3D bounding boxes during training=true2026.030.2—