The 3D semantic instance prediction task involves detecting and segmenting the object in an 3D scan mesh.

Evaluation and metrics

Similarly to the ScanNet benchmark in ScanNet200 our evaluation ranks all methods according to the average precision for each class. We report the mean average precision AP at overlap 0.25 (AP 25%), overlap 0.5 (AP 50%), and over overlaps in the range [0.5:0.95:0.05] (AP) for all 200 categories. Note that multiple predictions of the same ground truth instance are penalized as false positives.



This table lists the benchmark results for the ScanNet200 3D semantic instance scenario.




Method Infoavg ap 25%head ap 25%common ap 25%tail ap 25%alarm clockarmchairbackpackbagballbarbasketbathroom cabinetbathroom counterbathroom stallbathroom stall doorbathroom vanitybathtubbedbenchbicyclebinblackboardblanketblindsboardbookbookshelfbottlebowlboxbroombucketbulletin boardcabinetcalendarcandlecartcase of water bottlescd caseceilingceiling lightchairclockclosetcloset doorcloset rodcloset wallclothesclothes dryercoat rackcoffee kettlecoffee makercoffee tablecolumncomputer towercontainercopiercouchcountercratecupcurtaincushiondecorationdeskdining tabledish rackdishwasherdividerdoordoorframedresserdumbbelldustpanend tablefanfile cabinetfire alarmfire extinguisherfireplacefolded chairfurnitureguitarguitar casehair dryerhandicap barhatheadphonesironing boardjacketkeyboardkeyboard pianokitchen cabinetkitchen counterladderlamplaptoplaundry basketlaundry detergentlaundry hamperledgelightlight switchluggagemachinemailboxmatmattressmicrowavemini fridgemirrormonitormousemusic standnightstandobjectoffice chairottomanovenpaperpaper bagpaper cutterpaper towel dispenserpaper towel rollpersonpianopicturepillarpillowpipeplantplateplungerposterpotted plantpower outletpower stripprinterprojectorprojector screenpurserackradiatorrailrange hoodrecycling binrefrigeratorscaleseatshelfshoeshowershower curtainshower curtain rodshower doorshower floorshower headshower wallsignsinksoap dishsoap dispensersofa chairspeakerstair railstairsstandstoolstorage binstorage containerstorage organizerstovestructurestuffed animalsuitcasetabletelephonetissue boxtoastertoaster oventoilettoilet papertoilet paper dispensertoilet paper holdertoilet seat cover dispensertoweltrash bintrash cantraytubetvtv standvacuum cleanerventwardrobewashing machinewater bottlewater coolerwater pitcherwhiteboardwindowwindowsill
sort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
Mask3D Scannet2000.445 60.653 50.392 60.254 60.648 50.097 40.125 100.000 30.000 50.000 10.657 10.971 10.451 51.000 11.000 10.640 20.500 40.045 31.000 10.241 60.409 40.363 30.440 50.686 70.300 50.000 50.201 60.000 10.009 20.290 60.556 41.000 10.000 10.063 80.000 40.830 50.573 20.844 40.333 30.204 40.058 100.158 100.552 70.056 30.000 11.000 10.725 70.750 10.927 11.000 10.888 70.042 80.120 70.615 90.226 10.250 50.890 20.792 20.677 50.510 50.818 60.699 60.512 70.167 100.125 40.315 70.943 10.309 10.017 80.200 50.000 20.188 50.000 50.183 80.815 61.000 10.827 40.741 50.442 60.414 90.600 20.000 20.000 10.458 30.049 60.321 40.381 10.000 40.908 40.400 30.841 50.260 60.710 40.966 40.265 30.000 10.924 20.152 30.025 40.500 30.027 30.028 51.000 10.556 100.016 60.080 100.500 20.694 70.608 60.084 60.604 70.194 50.538 70.000 20.500 10.000 40.354 90.000 31.000 10.000 40.761 70.930 40.053 90.890 61.000 10.008 20.262 60.358 71.000 11.000 10.792 80.966 21.000 10.765 60.004 20.930 40.780 70.330 40.027 50.625 40.974 40.050 10.412 100.021 40.000 40.000 20.778 20.000 20.000 10.493 60.746 50.454 60.335 60.396 40.930 100.551 51.000 10.552 60.606 10.853 40.000 10.004 50.806 41.000 10.727 60.000 10.042 70.745 60.000 10.399 80.391 40.630 50.721 50.619 3
Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang, Bastian Leibe: Mask3D for 3D Semantic Instance Segmentation. ICRA 2023
CSC-Pretrain Inst.permissive0.275 100.466 100.218 90.110 100.625 70.007 100.500 10.000 30.000 50.000 10.000 100.222 100.377 91.000 10.661 100.400 60.000 90.000 70.000 70.119 100.000 50.000 100.277 90.685 80.067 80.000 50.132 80.000 10.000 30.000 90.367 90.000 70.000 10.000 90.000 40.591 80.055 90.783 90.000 80.014 80.500 50.161 90.278 80.000 50.000 10.667 60.768 50.500 20.866 61.000 10.829 90.000 90.019 100.555 100.000 40.000 80.305 100.000 70.750 10.200 90.783 90.429 80.395 80.677 30.020 100.286 80.584 100.000 40.000 90.115 100.000 20.000 70.000 50.145 100.423 100.500 70.364 100.369 90.571 40.448 80.206 100.000 20.000 10.200 70.106 40.065 100.000 50.000 40.750 70.200 70.774 60.000 100.501 80.841 90.000 50.000 10.692 100.063 90.000 80.000 80.000 50.000 80.500 90.649 70.000 70.084 90.125 90.719 50.413 100.004 90.450 100.000 70.638 40.000 20.000 80.000 40.505 80.000 30.000 70.000 40.727 80.833 60.221 50.779 90.000 70.000 50.168 100.311 100.125 60.571 90.500 90.143 100.000 50.250 90.000 30.869 70.667 90.162 100.000 70.250 91.000 10.000 50.500 60.000 70.000 40.000 20.689 90.000 20.000 10.312 90.383 100.114 80.333 70.000 90.997 60.420 70.613 90.212 100.500 20.819 70.000 10.000 60.768 71.000 10.918 30.000 10.000 80.278 100.000 10.333 90.000 100.353 70.546 100.258 9
Ji Hou, Benjamin Graham, Matthias Nießner, Saining Xie: Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts. CVPR 2021
TD3D Scannet200permissive0.379 70.603 70.306 70.190 70.635 60.073 60.500 10.000 30.000 50.000 10.495 60.735 40.275 101.000 10.979 60.590 30.000 90.021 50.000 70.146 80.000 50.356 40.173 100.795 30.226 70.000 50.173 70.000 10.000 30.226 70.390 60.000 70.000 10.250 50.000 40.706 70.061 80.885 30.093 40.186 50.259 90.200 60.667 40.000 50.000 10.667 60.825 30.250 70.834 91.000 10.958 10.553 10.111 80.748 60.220 20.051 70.866 50.792 20.390 100.045 100.800 70.302 100.517 60.533 40.113 60.427 50.843 50.000 40.458 50.600 10.000 20.101 60.000 50.259 50.717 70.500 70.615 70.520 60.526 50.457 60.270 90.000 20.000 10.400 40.088 50.294 50.181 30.000 41.000 10.400 30.710 100.103 80.477 100.905 60.061 40.000 10.906 40.102 60.232 20.125 70.000 50.003 70.792 81.000 10.000 70.102 80.125 90.559 100.523 80.075 70.715 50.000 70.424 100.000 20.396 40.250 10.638 60.000 30.000 70.000 40.622 100.833 60.221 40.970 10.250 50.038 10.260 70.415 60.125 61.000 11.000 10.857 40.000 50.908 10.012 10.869 80.836 50.635 10.111 40.625 41.000 10.020 30.510 50.003 60.009 31.000 10.778 20.000 20.000 10.370 80.755 40.288 70.333 70.274 61.000 10.557 40.731 60.456 70.433 30.769 100.000 10.000 60.621 91.000 10.458 90.000 10.196 40.817 30.000 10.472 40.222 70.205 100.689 60.274 8
Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: Top-Down Beats Bottom-Up in 3D Instance Segmentation. WACV 2024
CompetitorFormer-2000.469 40.676 40.401 50.296 40.692 40.057 70.500 10.083 20.000 50.000 10.534 40.701 50.410 70.903 90.998 30.878 10.500 40.068 10.250 60.424 11.000 10.244 60.556 10.696 50.270 61.000 10.240 50.000 10.000 30.587 30.380 71.000 10.000 10.500 10.000 40.900 30.257 40.901 10.085 60.207 30.863 10.224 51.000 10.109 20.000 10.724 50.806 40.500 20.869 51.000 10.829 80.247 30.474 10.759 50.021 30.269 40.873 30.125 60.467 90.542 10.885 40.829 20.711 30.285 90.118 50.482 40.770 90.025 30.018 70.400 20.000 20.677 20.500 10.222 60.916 11.000 10.818 50.827 20.342 80.650 40.452 40.000 20.000 10.330 60.173 20.278 60.000 50.083 31.000 10.336 60.748 70.508 50.698 50.989 20.286 20.000 10.933 10.175 20.400 10.663 20.015 40.103 41.000 10.829 20.125 40.293 50.500 20.847 40.711 30.295 20.543 90.385 30.581 60.000 20.500 10.000 40.747 40.050 21.000 10.013 10.850 40.886 50.214 70.918 40.125 60.000 50.320 50.610 30.025 100.933 81.000 10.820 60.250 40.901 20.000 30.980 10.878 10.325 50.160 30.574 60.703 80.009 40.540 40.011 50.000 40.000 20.700 80.056 10.000 10.491 70.729 60.617 50.489 50.565 21.000 10.410 80.750 50.629 50.292 40.839 60.000 10.157 10.839 21.000 10.834 50.000 10.131 50.794 40.000 10.667 10.144 80.664 30.854 10.500 5
ODIN - Ins200permissive0.451 50.637 60.407 40.277 50.583 90.116 30.500 10.000 30.125 20.000 10.599 20.823 20.407 80.667 100.941 70.542 41.000 10.000 71.000 10.162 70.000 50.028 90.357 60.695 60.550 10.000 50.475 20.000 10.000 30.714 10.626 21.000 10.000 10.500 10.125 20.749 60.080 60.742 100.528 20.078 70.500 50.334 20.667 40.333 10.000 10.278 100.723 80.250 70.859 81.000 10.826 100.108 70.221 50.763 40.000 40.250 50.742 70.500 50.750 10.400 70.855 50.769 40.701 40.469 60.203 30.406 60.870 30.000 40.963 10.200 50.000 20.000 70.500 10.370 30.886 41.000 10.782 60.504 70.429 70.494 50.337 70.000 20.000 10.600 10.000 80.215 70.226 20.000 40.944 30.200 70.887 20.750 10.874 10.877 70.438 10.000 10.867 50.089 70.003 60.500 30.000 50.333 21.000 10.742 60.125 40.671 30.417 80.616 90.637 50.238 40.873 20.528 20.494 90.000 20.250 50.000 40.688 50.000 31.000 10.000 40.872 20.833 60.275 20.779 91.000 10.000 50.441 30.577 50.167 41.000 10.500 90.777 70.000 50.778 50.000 30.910 60.800 60.232 80.019 60.717 20.833 50.000 50.638 20.284 10.000 40.000 20.778 20.000 20.000 10.597 40.699 70.850 20.333 70.250 70.944 80.571 20.677 70.795 40.264 50.852 50.000 10.000 60.824 31.000 10.668 70.000 10.000 80.667 70.000 10.333 90.333 50.760 10.679 70.404 6
DINO3D-Scannet200copyleft0.511 30.685 30.484 10.331 30.864 10.220 10.500 10.000 30.042 40.000 10.576 30.746 30.744 21.000 11.000 10.355 101.000 10.048 20.000 70.327 40.000 50.494 20.532 20.596 90.496 20.250 30.481 10.000 10.000 30.714 10.629 11.000 10.000 10.250 50.663 10.861 40.436 30.892 20.667 10.244 10.385 70.421 11.000 10.000 50.000 10.764 40.719 90.500 20.889 31.000 10.907 50.111 60.378 20.778 30.000 40.595 10.905 10.708 40.750 10.542 10.890 30.754 50.761 20.798 10.220 20.683 20.817 60.000 40.600 30.200 50.500 10.944 10.125 40.334 40.856 50.792 60.873 10.756 40.777 20.803 10.675 10.000 20.000 10.200 70.298 10.412 20.000 50.000 40.719 90.800 10.923 10.750 10.798 30.960 50.000 50.000 10.856 60.142 40.001 70.417 60.000 50.014 61.000 10.824 30.559 10.700 10.500 20.863 30.816 10.163 50.944 10.764 10.714 20.000 20.250 50.000 41.000 10.063 11.000 10.000 40.789 60.974 10.079 80.851 80.000 70.000 50.468 20.702 10.167 41.000 11.000 10.857 40.000 50.867 40.000 30.968 20.845 40.264 70.419 10.500 70.667 90.000 50.677 10.028 30.194 20.000 20.857 10.000 20.000 10.699 30.821 20.930 10.850 30.346 50.944 80.579 10.866 40.850 10.221 60.911 30.000 10.011 40.806 50.764 100.860 40.000 10.472 10.794 40.000 10.667 10.655 30.655 40.811 40.528 4
Jinyuan Qu, Hongyang Li, Xingyu Chen, Shilong Liu, Yukai Shi, Tianhe Ren, Ruitao Jing and Lei Zhang: SegDINO3D: 3D Instance Segmentation Empowered by Both Image-Level and Object-Level 2D Features. AAAI 2026
ACGP-ScanNet2000.544 10.737 10.483 20.381 10.747 30.184 20.500 10.500 10.708 10.000 10.371 90.604 60.718 31.000 10.996 40.511 51.000 10.035 41.000 10.388 30.528 30.307 50.517 30.834 10.405 41.000 10.396 30.000 10.042 10.540 40.536 51.000 10.000 10.500 10.050 30.918 20.189 50.801 80.086 50.140 60.667 30.292 31.000 10.000 50.000 10.903 20.838 10.444 60.902 21.000 10.912 40.213 40.367 30.858 10.000 40.396 20.846 61.000 10.677 50.443 60.921 20.905 10.573 50.500 50.064 80.658 30.931 20.144 20.637 20.317 40.000 20.638 40.000 50.462 10.897 31.000 10.860 20.944 10.759 30.682 30.510 30.000 20.000 10.391 50.140 30.462 10.143 40.250 20.884 50.600 20.851 40.677 30.680 60.993 10.000 50.000 10.825 70.185 10.036 31.000 10.089 10.472 11.000 10.765 50.500 30.671 30.500 20.929 10.699 40.349 10.738 40.204 40.782 10.008 10.146 70.024 30.835 30.000 30.250 60.006 30.854 30.974 10.590 10.968 21.000 10.007 30.483 10.666 21.000 11.000 11.000 10.959 31.000 10.867 30.000 30.955 30.850 30.403 20.341 20.637 30.721 70.000 50.552 30.000 70.000 40.000 20.778 20.000 20.000 10.759 10.859 10.767 31.000 10.944 11.000 10.424 60.944 20.840 20.206 70.915 20.000 10.064 20.800 61.000 11.000 10.000 10.461 20.903 20.000 10.507 30.903 10.718 20.842 20.809 1
Volt-SPFormerpermissive0.527 20.731 20.475 30.342 20.789 20.076 50.500 10.000 30.125 20.000 10.391 80.508 90.753 11.000 10.994 50.400 60.500 40.020 61.000 10.413 20.850 20.547 10.510 40.810 20.433 30.250 30.346 40.000 10.000 30.519 50.594 31.000 10.000 10.331 40.000 40.937 10.638 10.826 50.056 70.214 20.850 20.262 40.667 40.028 40.000 10.817 30.825 20.250 70.880 41.000 10.950 20.279 20.309 40.856 20.000 40.304 30.867 40.000 70.750 10.542 10.942 10.818 30.901 10.458 70.329 10.750 10.855 40.000 40.510 40.200 50.000 20.677 20.500 10.397 20.903 21.000 10.843 30.773 31.000 10.799 20.449 50.250 10.000 10.600 10.027 70.372 30.000 51.000 10.833 60.400 30.878 30.656 40.843 20.973 30.000 50.000 10.921 30.103 50.008 50.500 30.057 20.278 31.000 10.802 40.557 20.700 11.000 10.874 20.767 20.279 30.801 30.047 60.714 20.000 20.500 10.250 10.907 20.000 31.000 10.011 20.875 10.944 30.255 30.923 31.000 10.002 40.321 40.579 41.000 11.000 11.000 11.000 11.000 10.737 70.000 30.926 50.857 20.343 30.000 70.741 10.629 100.025 20.500 60.000 70.000 40.000 20.725 70.000 20.000 10.715 20.803 30.738 41.000 10.500 31.000 10.565 30.884 30.812 30.167 80.937 10.000 10.019 30.923 11.000 11.000 10.000 10.099 61.000 10.000 10.472 40.764 20.614 60.815 30.681 2
Kadir Yilmaz, Adrian Kruse, Tristan Höfer, Daan de Geus, Bastian Leibe: Volume Transformer: Revisiting Vanilla Transformers for 3D Scene Understanding.
Minkowski 34D Inst.permissive0.280 90.488 90.192 100.124 90.593 80.010 90.500 10.000 30.000 50.000 10.447 70.535 80.445 61.000 10.861 90.400 60.225 70.000 70.000 70.142 90.000 50.074 80.342 80.467 100.067 80.000 50.119 100.000 10.000 30.000 90.337 100.000 70.000 10.000 90.000 40.506 100.070 70.804 70.000 80.000 90.333 80.172 80.150 100.000 50.000 10.479 90.745 60.000 100.830 101.000 10.904 60.167 50.090 90.732 70.000 40.000 80.443 90.000 70.500 70.542 10.772 100.396 90.077 100.385 80.044 90.118 100.777 80.000 40.000 90.200 50.000 20.000 70.000 50.148 90.502 90.500 70.419 90.159 100.281 90.404 100.317 80.000 20.000 10.200 70.000 80.077 80.000 50.000 40.750 70.200 70.715 90.021 90.551 70.828 100.000 50.000 10.743 90.059 100.000 80.000 80.000 50.000 80.125 100.648 80.000 70.191 70.500 20.669 80.502 90.000 100.568 80.000 70.516 80.000 20.000 80.000 40.305 100.000 30.000 70.000 40.825 50.833 60.021 100.918 40.000 70.000 50.191 90.346 90.100 80.981 71.000 10.286 90.000 50.000 100.000 30.868 90.648 100.292 60.000 70.375 81.000 10.000 50.500 60.000 70.333 10.000 20.538 100.000 20.000 10.213 100.518 90.098 90.528 40.250 70.997 60.284 100.677 70.398 80.167 80.790 90.000 10.000 60.618 100.903 90.200 100.000 10.333 30.333 90.000 10.442 70.083 90.213 90.587 90.131 10
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
LGround Inst.permissive0.314 80.529 80.225 80.155 80.578 100.010 80.500 10.000 30.000 50.000 10.515 50.556 70.696 41.000 10.927 80.400 60.083 80.000 71.000 10.252 50.000 50.167 70.350 70.731 40.067 80.000 50.123 90.000 10.000 30.036 80.372 80.000 70.000 10.250 50.000 40.569 90.031 100.810 60.000 80.000 90.630 40.183 70.278 80.000 50.000 10.582 80.589 100.500 20.863 71.000 10.940 30.000 90.144 60.716 80.000 40.000 80.484 80.000 70.500 70.400 70.798 80.500 70.278 90.750 20.093 70.166 90.783 70.000 40.200 60.400 20.000 20.000 70.000 50.219 70.539 80.500 70.578 80.413 80.181 100.457 70.375 60.000 20.000 10.050 100.000 80.077 90.000 50.000 40.500 100.000 100.743 80.250 70.488 90.846 80.000 50.000 10.800 80.069 80.000 80.000 80.000 50.000 81.000 10.607 90.000 70.200 60.500 20.694 60.528 70.063 80.659 60.000 70.594 50.000 20.000 80.000 40.571 70.000 30.000 70.000 40.716 90.647 100.221 50.857 70.000 70.000 50.217 80.346 80.071 90.530 101.000 10.429 80.000 50.286 80.000 30.826 100.706 80.208 90.000 70.250 90.744 60.000 50.500 60.042 20.000 40.000 20.746 60.000 20.000 10.517 50.625 80.085 100.333 70.000 91.000 10.378 90.533 100.376 90.042 100.814 80.000 10.000 60.765 81.000 10.600 80.000 10.000 80.667 70.000 10.472 40.333 50.337 80.605 80.305 7
David Rozenberszki, Or Litany, Angela Dai: Language-Grounded Indoor 3D Semantic Segmentation in the Wild.