3D Instance Segmentation on S3DIS (6-fold CV)
84.5Mean Precision @50% IoURPGN
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| RPGN2023.03 | 84.5 | — | 70.5 | — | — | — | — | — | |
| RPGNEvaluation Split=6-fold cross-validation2022.07 | 84.5 | — | 70.5 | — | — | — | — | — | |
| OneFormer3D2023.11 | 82.3 | 75.8 | 74.1 | — | — | — | 63 | — | |
| PBNet2023.11 | 80.1 | 70.6 | 72.9 | — | — | — | 59.5 | — | |
| PBNetEvaluation Split=6-fold cross-validation2022.07 | 80.1 | 59.5 | 72.9 | — | — | — | — | — | |
| ISBNet2023.03 | 77.5 | 70.5 | 77.1 | 60.8 | 74.9 | 76.8 | — | — | |
| Mask3DPre-trained on ScanNet=true, Fine-tuned on S3DIS=true2022.10 | 76.5 | 74.3 | 66.2 | 61.8 | — | — | — | — | |
| Mask3D2023.11 | 76.5 | 74.3 | 66.2 | — | — | — | 61.8 | — | |
| PointInst3D2023.03 | 76.4 | — | 74 | — | 71.5 | 74.1 | — | — | |
| PointInst3DEvaluation Split=6-fold cross-validation2022.07 | 76.4 | — | 74 | — | — | — | — | — | |
| TD3D2023.11 | 76.3 | 68.2 | 74 | — | — | — | 56.2 | — | |
| SoftGroup++subsampling=none2022.09 | 75.9 | 71.3 | 74.4 | 56.6 | — | — | — | — | |
| SoftGroupEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 75.3 | — | 69.8 | — | 69.3 | 71.7 | — | — | |
| DKNetEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 75.3 | — | 71.1 | — | 70.3 | 72.8 | — | — | |
| SoftGroupevaluation_split=6-fold cross-validation2022.11 | 75.3 | 68.9 | 69.8 | — | — | — | — | — | |
| SoftGroup2023.03 | 75.3 | 68.9 | 69.8 | 54.4 | 69.3 | 71.7 | — | — | |
| DKNet2023.03 | 75.3 | — | 71.1 | — | 70.3 | 72.8 | — | — | |
| SoftGroupPre-trained on ScanNet=true, Fine-tuned on S3DIS=true2022.10 | 75.3 | 68.9 | 69.8 | 54.4 | — | — | — | — | |
| DKNet2023.11 | 75.3 | — | 71.1 | — | — | — | — | — | |
| SoftGroup2023.11 | 75.3 | 68.9 | 69.8 | — | — | — | 54.4 | — | |
| SoftGroupsubsampling=1/42022.09 | 75.3 | 68.9 | 69.8 | 54.4 | — | — | — | — | |
| SoftGroupEvaluation Split=6-fold cross-validation2022.07 | 75.3 | 54.4 | 69.8 | — | — | — | — | — | |
| DKNetEvaluation Split=6-fold cross-validation2022.07 | 75.3 | — | 71.1 | — | — | — | — | — | |
| SPFormerevaluation_split=6-fold cross-validation2022.11 | 74 | 69.2 | 71.1 | — | — | — | — | — | |
| SPFormer2023.11 | 74 | 69.2 | 71.1 | — | — | — | — | — | |
| SoftGroupevaluation_protocol=6-fold cross validation2022.03 | 73.6 | 66.1 | 66.6 | 51.6 | 66.1 | 68 | — | — | |
| SSTNetEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 73.5 | — | 73.4 | — | — | — | — | — | |
| SSTNetevaluation_split=6-fold cross-validation2022.11 | 73.5 | 67.8 | 73.4 | — | — | — | — | — | |
| SSTNet2023.03 | 73.5 | 67.8 | 73.4 | 54.1 | — | — | — | — | |
| SSTNet2022.10 | 73.5 | 67.8 | 73.4 | 54.1 | — | — | — | — | |
| SSTNet2023.11 | 73.5 | 67.8 | 73.4 | — | — | — | 54.1 | — | |
| SSTNet2022.09 | 73.5 | 67.8 | 73.4 | 54.1 | — | — | — | — | |
| SSTNetevaluation_protocol=6-fold cross validation2021.08 | 73.5 | 67.8 | 73.4 | — | — | — | 54.1 | — | |
| SSTNetEvaluation Split=6-fold cross-validation2022.07 | 73.5 | 54.1 | 73.4 | — | — | — | — | — | |
| HAISEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 73.2 | — | 69.4 | — | 67 | 70.4 | — | — | |
| HAISevaluation_split=6-fold cross-validation2022.11 | 73.2 | — | 69.4 | — | — | — | — | — | |
| HAIS2023.03 | 73.2 | — | 69.4 | — | 67 | 70.4 | — | — | |
| HAIS2023.11 | 73.2 | — | 69.4 | — | — | — | — | — | |
| HAIS2022.09 | 73.2 | — | 69.4 | — | — | — | — | — | |
| HAISEvaluation Split=6-fold cross-validation2022.07 | 73.2 | — | 69.4 | — | — | — | — | — | |
| OccuSegEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 72.8 | — | 60.3 | — | — | — | — | — | |
| OccuSeg2023.03 | 72.8 | — | 60.3 | — | — | — | — | — | |
| Mask3D2022.10 | 72.8 | 75.5 | 74.5 | 64.5 | — | — | — | — | |
| OccuSegevaluation_protocol=6-fold cross validation2021.08 | 72.8 | — | 60.3 | — | — | — | — | — | |
| HAISevaluation_protocol=6-fold cross validation2022.03 | 71.1 | — | 65 | — | 64.3 | 66 | — | — | |
| PointGroupEvaluation Protocol=6-fold cross validation2020.04 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroupEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 69.6 | — | 69.2 | — | — | — | — | — | |
| PointGroup2022.03 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroupevaluation_split=6-fold cross-validation2022.11 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroup2023.03 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| MaskGroup2022.10 | 69.6 | 66.6 | — | 69.9 | — | — | — | — | |
| PointGroup2023.11 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroup2022.09 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroupevaluation_protocol=6-fold cross validation2021.08 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| PointGroupEvaluation Split=6-fold cross-validation2022.07 | 69.6 | 64 | 69.2 | — | — | — | — | — | |
| HAISPre-trained on ScanNet=true, Fine-tuned on S3DIS=true2022.10 | 69.4 | 73.2 | — | — | — | — | — | — | |
| PointGroup2022.10 | 69.2 | 69.6 | — | 64 | — | — | — | — | |
| GICN2020.07 | 68.5 | — | 50.8 | — | — | — | — | — | |
| GICNEvaluation Protocol=6-fold cross validation, Voxel Size=0.05m2022.07 | 68.5 | — | 50.8 | — | — | — | — | — | |
| GICNevaluation_split=6-fold cross-validation2022.11 | 68.5 | — | 50.8 | — | — | — | — | — | |
| MaskGroup2022.03 | 66.6 | 69.9 | 69.6 | — | — | — | — | — | |
| MaskGroupEvaluation Split=6-fold cross-validation2022.07 | 66.6 | 69.9 | 69.6 | — | — | — | — | — | |
| 3D-BoNetEvaluation Protocol=6-fold cross validation2020.04 | 65.6 | — | 47.6 | — | — | — | — | — | |
| 3D-BoNet2020.07 | 65.6 | — | 47.6 | — | — | — | — | — | |
| 3D-BEVIS2020.07 | 65.6 | — | — | — | — | — | — | — | |
| 3D-BoNetevaluation_split=6-fold cross-validation2022.11 | 65.6 | — | 47.7 | — | — | — | — | — | |
| 3D-BoNet2023.03 | 65.6 | — | 47.7 | — | — | — | — | — | |
| 3D-BoNet2022.09 | 65.6 | — | 47.7 | — | — | — | — | — | |
| 3D-BoNetevaluation_protocol=6-fold cross validation2021.08 | 65.6 | — | 47.6 | — | — | — | — | — | |
| SSTNetevaluation_protocol=6-fold cross validation2022.03 | 65.5 | 59.3 | 64.2 | 42.7 | — | — | — | — | |
| 3D-MPA2022.10 | 64.1 | 66.7 | — | — | — | — | — | — | |
| ASISEvaluation Protocol=6-fold cross validation2020.04 | 63.6 | — | 47.5 | — | — | — | — | — | |
| ASIS2020.07 | 63.6 | — | 47.5 | — | — | — | — | — | |
| ASIS2022.09 | 63.6 | — | 47.5 | — | — | — | — | — | |
| ASISevaluation_protocol=6-fold cross validation2021.08 | 63.6 | — | 47.5 | — | — | — | — | — | |
| PointGroupevaluation_protocol=6-fold cross validation2022.03 | 61.9 | 57.8 | 62.1 | — | — | — | — | — | |
| OccuSeg2022.03 | 60.3 | 72.8 | — | — | — | — | — | — | |
| OccuSeg2022.10 | 60.3 | 72.8 | — | — | — | — | — | — | |
| PartNetEvaluation Protocol=6-fold cross validation2020.04 | 56.4 | — | 43.4 | — | — | — | — | — | |
| ASISevaluation_protocol=6-fold cross validation2022.03 | 55.3 | — | 42.4 | — | 44.6 | 47.8 | — | — | |
| MPNet2022.03 | 53.7 | 68.4 | — | — | — | — | — | — | |
| InsEmb2022.03 | 51.8 | 67.2 | — | — | — | — | — | — | |
| GICN2022.03 | 50.8 | 68.5 | — | — | — | — | — | — | |
| ICM-3D2022.03 | 49.8 | 65.9 | — | — | — | — | — | — | |
| 3D-Bonet2022.10 | 47.6 | 65.6 | — | — | — | — | — | — | |
| ASIS2022.10 | 47.5 | 63.6 | — | — | — | — | — | — | |
| SGPNEvaluation Protocol=6-fold cross validation2020.04 | 38.2 | 54.4 | 31.2 | — | — | — | — | — | |
| SGPN2023.03 | 38.2 | 54.4 | 31.2 | — | 37.9 | 40.8 | — | — | |
| SGPN2022.09 | 38.2 | — | 31.2 | — | — | — | — | — | |
| SGPNEvaluation Split=6-fold cross-validation2022.07 | 38.2 | — | 31.2 | — | — | — | — | — | |
| SGPNevaluation_protocol=6-fold cross validation2022.03 | 36 | — | 28.7 | — | 32.7 | 35.5 | — | — | |
| SGPN2022.10 | 31.2 | 38.2 | — | — | — | — | — | — | |
| 3D-BoNet2020.03 | — | 65.6 | 47.6 | — | — | — | — | — | |
| 3D-MPA2020.03 | — | 66.7 | 64.1 | — | — | — | — | — | |
| ASIS2020.03 | — | 63.6 | 47.5 | — | — | — | — | — | |
| EFEMLearning Paradigm=Unsupervised2026.05 | — | 37.8 | — | 16.2 | — | — | — | 45.9 | |
| EvObj (Ours-Diffusion)Learning Paradigm=Unsupervised, Model Type=Diffusion2026.05 | — | 69.3 | — | 47.4 | — | — | — | 78.6 | |
| EvObj (Ours-VAE)Learning Paradigm=Unsupervised, Model Type=VAE2026.05 | — | 67.5 | — | 46.3 | — | — | — | 77.4 | |
| FoundObjSupervision Level=Unsupervised2026.05 | — | 24 | — | 11.4 | — | — | — | 45.7 | |
| GrabSSupervision Level=Unsupervised2026.05 | — | 5.5 | — | 3.2 | — | — | — | 9.4 |