ScanNet200 3D Semantic Instance Benchmark

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 50%head ap 50%common ap 50%tail ap 50%chairtabledoorcouchcabinetshelfdeskoffice chairbedpillowsinkpicturewindowtoiletbookshelfmonitorcurtainbookarmchaircoffee tableboxrefrigeratorlampkitchen cabinettowelclothestvnightstandcounterdresserstoolcushionplantceilingbathtubend tabledining tablekeyboardbagbackpacktoilet paperprintertv standwhiteboardblanketshower curtaintrash canclosetstairsmicrowavestoveshoecomputer towerbottlebinottomanbenchboardwashing machinemirrorcopierbasketsofa chairfile cabinetfanlaptopshowerpaperpersonpaper towel dispenserovenblindsrackplateblackboardpianosuitcaserailradiatorrecycling bincontainerwardrobesoap dispensertelephonebucketclockstandlightlaundry basketpipeclothes dryerguitartoilet paper holderseatspeakercolumnbicycleladderbathroom stallshower wallcupjacketstorage bincoffee makerdishwasherpaper towel rollmachinematwindowsillbartoasterbulletin boardironing boardfireplacesoap dishkitchen counterdoorframetoilet paper dispensermini fridgefire extinguisherballhatshower curtain rodwater coolerpaper cuttertrayshower doorpillarledgetoaster ovenmousetoilet seat cover dispenserfurniturecartstorage containerscaletissue boxlight switchcratepower outletdecorationsignprojectorcloset doorvacuum cleanercandleplungerstuffed animalheadphonesdish rackbroomguitar caserange hooddustpanhair dryerwater bottlehandicap barpurseventshower floorwater pitchermailboxbowlpaper bagalarm clockmusic standprojector screendividerlaundry detergentbathroom counterobjectbathroom vanitycloset walllaundry hamperbathroom stall doorceiling lighttrash bindumbbellstair railtubebathroom cabinetcd casecloset rodcoffee kettlestructureshower headkeyboard pianocase of water bottlescoat rackstorage organizerfolded chairfire alarmpower stripcalendarposterpotted plantluggagemattress
sort 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 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 by
AQ3D-ScanNet2000.491 20.675 10.458 10.308 30.907 10.712 20.846 30.860 40.596 10.305 40.796 10.328 20.955 40.768 20.662 60.925 10.666 21.000 10.754 30.919 50.824 10.459 10.804 50.307 30.929 20.698 30.648 20.737 30.310 10.764 50.896 20.049 60.778 20.576 30.160 30.879 20.832 41.000 10.077 70.023 70.829 30.153 30.823 20.517 50.667 61.000 10.498 60.371 41.000 10.934 10.005 71.000 10.831 50.778 20.551 30.754 110.258 60.025 50.564 40.458 31.000 10.521 31.000 10.833 10.347 30.991 10.473 50.434 61.000 10.071 20.428 50.868 21.000 10.333 50.133 10.000 70.533 51.000 10.671 30.765 30.534 20.664 10.818 50.698 30.083 10.009 30.000 50.431 10.000 20.147 10.667 20.029 20.172 30.250 61.000 10.737 30.104 50.713 10.362 20.319 30.004 20.489 90.600 10.083 50.563 20.400 30.923 10.500 10.716 10.605 30.500 10.506 50.530 30.750 30.658 40.318 60.500 10.000 21.000 10.333 81.000 10.000 10.105 30.000 40.500 10.000 31.000 11.000 10.000 50.141 80.026 80.000 30.000 31.000 10.000 10.640 20.000 10.517 40.750 10.000 11.000 10.000 10.000 10.714 70.667 30.479 30.167 50.667 10.000 11.000 10.000 10.074 70.834 10.517 30.444 70.073 60.433 30.438 60.281 20.083 20.000 10.553 10.000 10.000 21.000 10.917 60.000 30.056 20.000 20.200 40.141 30.000 40.250 70.000 2
CompetitorFormer-2000.415 50.574 50.370 60.274 50.885 20.584 80.846 40.779 90.318 80.205 60.704 40.400 10.987 20.651 40.731 50.830 40.682 11.000 10.599 80.957 20.685 60.428 30.806 40.196 60.870 70.641 50.600 50.583 60.183 70.780 40.833 40.095 40.663 40.538 50.021 40.540 90.845 20.903 80.103 50.083 60.881 20.054 80.632 50.311 90.745 31.000 10.545 40.378 20.933 90.832 70.015 40.684 70.748 70.700 50.562 20.869 20.218 70.064 10.885 10.243 70.794 50.484 51.000 10.000 60.289 50.758 50.482 40.452 40.000 90.015 50.286 60.759 60.663 31.000 10.000 30.380 30.250 60.500 30.491 60.622 80.213 60.131 50.877 10.602 60.000 40.005 40.008 40.209 70.000 20.089 30.399 50.000 60.160 40.500 20.500 50.144 100.260 20.347 80.443 10.207 70.000 30.724 50.400 20.125 40.083 40.317 70.462 40.083 30.565 20.587 40.500 10.648 20.551 10.750 30.508 60.018 80.500 10.000 21.000 10.667 11.000 10.000 10.142 20.000 40.500 10.000 30.125 70.489 50.000 50.500 10.269 40.000 30.050 20.625 60.000 10.581 40.000 10.677 20.467 90.000 10.694 70.000 10.000 10.820 60.071 100.215 81.000 10.103 50.000 11.000 10.000 10.132 20.410 60.327 70.541 50.232 40.292 50.261 90.186 50.157 10.000 10.216 60.000 10.056 10.250 51.000 10.000 30.082 10.000 20.400 20.025 40.000 41.000 10.000 2
TD3D Scannet200permissive0.320 80.501 80.264 80.164 80.841 40.679 40.716 80.879 30.280 90.192 70.634 60.231 30.733 90.459 80.565 90.498 100.560 71.000 10.686 40.890 70.708 30.123 100.820 30.152 80.967 10.456 60.458 80.387 80.194 60.435 110.906 10.077 50.396 80.509 60.217 20.715 60.619 71.000 10.099 60.792 20.513 80.062 70.506 80.549 30.605 71.000 10.123 100.106 61.000 10.744 100.000 81.000 10.504 110.525 80.185 80.790 80.101 80.008 70.587 30.356 40.817 40.083 111.000 10.000 60.621 10.842 30.415 60.268 100.083 80.000 60.098 90.881 10.125 80.000 60.000 30.000 70.000 80.125 100.332 90.448 110.202 70.196 40.798 70.264 80.000 40.000 50.017 30.233 60.000 20.063 40.333 60.038 10.111 50.250 60.000 80.516 50.208 30.470 60.094 90.218 60.000 30.667 60.033 110.000 80.000 50.400 30.156 80.000 40.267 60.226 70.000 60.104 80.159 80.299 110.095 90.458 40.500 10.000 21.000 10.472 40.792 90.000 10.022 60.061 30.250 60.008 20.250 60.333 70.143 30.396 40.049 70.012 10.000 30.283 100.000 10.241 90.000 10.101 70.331 100.000 10.629 90.000 10.000 10.857 20.222 80.677 10.000 60.003 80.000 10.000 80.000 10.076 60.252 90.400 40.431 80.061 90.328 40.331 80.500 10.000 70.000 10.167 70.000 10.000 20.000 60.500 80.000 30.000 51.000 10.542 10.000 50.063 20.000 80.000 2
Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: Top-Down Beats Bottom-Up in 3D Instance Segmentation. WACV 2024
LGround Inst.permissive0.246 90.413 90.170 90.130 90.754 80.541 90.682 100.903 10.264 100.164 80.234 90.000 90.681 100.452 90.464 110.541 90.399 91.000 10.637 60.772 90.588 90.190 80.589 110.081 90.857 80.426 80.373 90.318 90.135 80.690 70.653 100.000 90.159 100.500 70.000 50.581 80.387 101.000 10.046 90.000 80.402 90.003 110.455 110.196 100.571 81.000 10.270 90.003 110.530 110.748 90.000 80.744 50.575 90.511 90.112 90.815 50.067 90.000 80.400 70.167 80.667 70.241 81.000 10.000 60.208 90.660 80.125 100.317 80.000 90.000 60.100 80.561 100.000 90.000 60.000 30.000 71.000 10.500 30.344 80.568 100.167 80.000 80.706 90.068 90.000 40.000 50.000 50.063 90.000 20.000 90.056 100.000 60.000 70.500 20.000 80.143 110.017 100.125 90.097 80.164 90.000 30.582 80.400 20.000 80.000 50.000 100.083 100.000 40.000 90.000 90.000 60.025 90.156 90.533 80.250 80.200 70.500 10.000 21.000 10.333 81.000 10.000 10.000 80.000 40.000 90.000 30.000 80.333 70.000 50.000 90.000 90.000 30.000 30.400 90.000 10.364 70.000 10.000 80.500 80.000 10.511 100.000 10.000 10.286 90.333 60.000 100.000 60.000 90.000 10.000 80.000 10.034 90.111 110.000 90.333 100.031 110.000 100.750 10.125 70.000 70.000 10.151 80.000 10.000 20.000 60.500 80.000 30.000 50.000 20.000 110.000 50.000 40.000 80.000 2
David Rozenberszki, Or Litany, Angela Dai: Language-Grounded Indoor 3D Semantic Segmentation in the Wild.
ACGP-ScanNet2000.494 10.656 20.453 30.345 10.771 70.746 10.864 20.831 80.484 60.344 10.687 50.035 40.933 60.633 50.823 10.871 20.633 30.997 50.834 10.963 10.718 20.442 20.838 10.313 20.893 50.680 40.655 10.840 10.276 30.800 20.743 80.212 10.573 60.548 40.000 50.736 40.853 11.000 10.062 81.000 10.819 50.178 20.723 40.424 70.835 21.000 10.618 20.378 31.000 10.914 30.023 30.703 60.836 40.667 60.495 50.817 40.260 50.035 40.511 60.307 60.903 30.566 21.000 10.708 20.360 20.931 20.511 30.510 31.000 10.089 10.671 10.765 51.000 10.528 30.005 20.204 41.000 10.500 30.702 20.842 20.590 10.461 20.844 40.664 40.042 20.010 20.000 50.405 20.000 20.040 70.817 10.007 40.341 20.444 51.000 10.714 40.020 90.643 20.310 40.396 10.000 30.903 20.400 20.500 20.250 30.600 20.657 20.500 10.304 50.540 50.000 60.532 30.545 20.750 30.677 30.637 20.500 10.000 20.792 70.504 31.000 10.000 10.071 40.000 40.396 40.000 31.000 11.000 10.036 40.146 70.396 20.000 30.000 31.000 10.000 10.627 30.000 10.638 30.677 60.000 10.974 20.000 10.000 10.835 50.850 10.449 51.000 10.472 20.000 10.250 70.000 10.148 10.695 40.387 60.592 30.189 50.204 80.500 40.088 90.064 30.000 10.471 20.000 10.000 21.000 11.000 10.000 30.000 50.000 20.317 30.144 20.024 31.000 10.008 1
Rongkun Yang, Ye Zhang, Longguang Wang, Zhiheng Fu, Lian Xu, Yulan Guo: Beyond Context Bias: Adaptive Instance Placement for Robust 3D Instance Segmentation.
Minkowski 34D Inst.permissive0.203 110.369 100.134 110.078 110.706 90.382 100.693 90.845 60.221 110.150 100.158 100.000 90.746 80.369 100.545 100.595 70.387 100.997 50.413 110.720 110.636 70.165 90.732 70.070 100.851 90.402 90.251 100.313 100.123 100.583 90.696 90.000 90.051 110.500 70.000 50.500 100.372 110.667 90.009 100.000 80.307 110.003 100.479 100.107 110.226 110.903 90.109 110.031 90.981 80.726 110.000 80.522 110.669 80.282 110.052 110.778 100.000 100.000 80.400 70.074 90.333 100.218 101.000 10.000 60.250 80.406 110.118 110.317 80.100 70.000 60.191 70.596 80.000 90.000 60.000 30.000 70.000 80.500 30.178 110.701 60.000 110.000 80.522 110.018 110.000 40.000 50.000 50.060 100.000 20.000 90.033 110.000 60.000 70.000 100.000 80.281 80.100 60.000 110.090 100.133 100.000 30.422 100.050 100.000 80.000 50.200 80.000 110.000 40.000 90.000 90.000 60.000 100.123 100.677 70.021 100.000 100.500 10.000 20.500 90.442 60.125 110.000 10.000 80.000 40.000 90.000 30.000 80.056 100.000 50.000 90.000 90.000 30.000 30.200 110.000 10.143 110.000 10.000 80.250 110.000 10.511 100.000 10.000 10.286 90.083 90.396 60.000 60.000 90.000 10.000 80.000 10.025 100.300 80.000 90.371 90.070 70.000 100.385 70.000 110.000 70.000 10.000 110.000 10.000 20.000 60.500 80.000 30.000 50.000 20.200 40.000 50.000 40.000 80.000 2
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
CSC-Pretrain Inst.permissive0.209 100.361 110.157 100.085 100.700 100.248 110.634 110.776 100.322 70.135 110.103 110.000 90.524 110.364 110.618 70.592 80.381 110.997 50.589 90.747 100.340 110.109 110.768 60.059 110.702 100.448 70.188 110.149 110.091 110.636 80.573 110.000 90.246 90.500 70.000 50.450 110.405 90.667 90.006 110.000 80.356 100.007 90.506 70.420 80.340 90.667 110.294 80.004 100.571 100.748 80.000 81.000 10.573 100.502 100.094 100.807 60.000 100.000 80.400 70.000 110.278 110.228 91.000 10.000 60.115 110.432 100.198 90.050 110.125 50.000 60.000 110.573 90.000 90.000 60.000 30.000 70.000 80.125 100.312 100.610 90.221 50.000 80.667 100.050 100.000 40.000 50.000 50.032 110.000 20.000 90.083 80.000 60.000 70.000 100.000 80.220 90.000 110.125 90.000 110.111 110.000 30.667 60.200 80.000 80.000 50.000 100.110 90.000 40.000 90.000 90.000 60.000 100.053 110.500 90.000 110.000 100.500 10.000 20.500 90.333 80.500 100.000 10.000 80.000 40.000 90.000 30.000 80.000 110.000 50.000 90.000 90.000 30.000 30.600 70.000 10.364 70.000 10.000 80.750 10.000 10.833 50.000 10.000 10.143 110.000 110.396 60.000 60.000 90.000 10.000 80.000 10.021 110.221 100.000 90.093 110.055 100.451 20.677 30.125 70.000 70.000 10.028 100.000 10.000 20.000 60.500 80.000 30.000 50.000 20.050 100.000 50.000 40.000 80.000 2
Ji Hou, Benjamin Graham, Matthias Nießner, Saining Xie: Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts. CVPR 2021
Volt-SPFormerpermissive0.475 30.630 30.451 40.314 20.806 60.666 60.923 10.847 50.541 40.224 50.755 20.008 60.994 10.735 30.818 20.869 30.621 40.990 80.811 20.894 60.702 40.423 40.825 20.281 40.923 30.787 20.564 60.699 40.245 40.784 30.800 50.129 20.900 10.500 70.000 50.768 30.841 31.000 10.319 10.000 80.903 10.068 60.772 30.565 20.683 51.000 10.546 30.410 11.000 10.930 20.014 50.629 100.878 20.725 40.499 40.799 70.412 30.019 60.400 70.500 11.000 10.612 11.000 10.125 30.343 40.823 40.750 10.449 50.250 40.056 30.585 30.797 40.500 40.667 20.000 30.043 61.000 11.000 10.716 10.853 10.255 40.099 60.857 20.651 50.000 40.000 50.025 20.375 30.250 10.056 50.333 60.002 50.000 70.250 60.500 51.000 10.107 40.613 30.294 50.300 50.000 30.817 30.400 20.500 21.000 10.400 30.452 50.000 40.500 30.519 60.500 10.372 60.482 40.750 30.641 50.510 30.500 10.000 21.000 10.472 41.000 10.000 10.026 50.000 40.331 50.000 31.000 11.000 10.000 50.500 10.304 30.000 30.000 31.000 10.000 10.714 10.000 10.500 50.750 10.000 10.944 40.000 10.000 10.857 20.764 20.455 40.250 30.278 40.000 11.000 10.000 10.078 50.742 20.600 10.524 60.638 10.167 90.208 100.209 40.019 40.000 10.241 50.000 10.000 21.000 11.000 10.000 30.028 40.000 20.200 40.000 50.250 11.000 10.000 2
Kadir Yilmaz, Adrian Kruse, Tristan Höfer, Daan de Geus, Bastian Leibe: Volume Transformer: Revisiting Vanilla Transformers for 3D Scene Understanding.
DINO3D-Scannet200copyleft0.454 40.587 40.453 20.296 40.871 30.703 30.845 50.891 20.572 20.312 30.753 30.001 80.981 30.773 10.767 30.771 50.614 50.944 90.586 100.937 30.690 50.381 50.716 100.409 10.918 40.803 10.602 40.777 20.290 20.721 60.779 70.096 30.728 30.677 10.000 50.944 10.793 51.000 10.214 20.708 40.823 40.200 10.851 10.499 61.000 10.764 100.473 70.248 51.000 10.911 40.216 10.667 80.824 60.857 10.616 10.842 30.496 20.046 20.355 110.494 20.405 90.507 41.000 10.042 50.264 70.743 60.683 20.675 10.125 50.000 60.600 20.816 30.417 70.000 60.000 30.764 10.000 80.500 30.563 50.720 50.079 90.442 30.845 30.835 20.000 40.000 50.000 50.324 40.000 20.117 20.083 80.000 60.419 10.500 21.000 10.777 20.378 10.594 40.361 30.327 20.000 30.764 40.400 20.548 10.000 50.800 10.437 60.000 40.346 40.714 10.125 50.662 10.475 50.866 20.750 10.400 50.500 10.500 11.000 10.667 11.000 10.000 10.298 10.000 40.250 60.194 10.000 80.850 40.000 50.250 50.595 10.000 30.063 10.520 80.000 10.571 50.000 10.944 10.750 10.000 10.974 20.000 10.000 10.857 20.655 40.000 100.250 30.014 70.000 11.000 10.000 10.116 30.729 30.200 80.545 40.436 30.221 70.750 10.177 60.011 50.000 10.284 30.000 10.000 20.000 60.792 70.050 20.000 50.000 20.200 40.000 50.000 41.000 10.000 2
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
ODIN - Ins200permissive0.381 70.507 70.375 50.237 60.653 110.614 70.780 60.744 110.566 30.328 20.446 80.003 70.853 70.496 70.582 80.448 110.434 80.938 100.682 50.782 80.494 100.274 70.723 90.269 50.694 110.393 100.511 70.695 50.227 50.550 100.795 60.039 70.602 50.638 20.000 50.734 50.585 80.667 90.163 30.500 50.769 60.108 40.484 90.569 10.688 41.000 10.665 10.093 71.000 10.863 50.049 20.667 80.887 10.778 20.422 60.786 90.550 10.000 80.542 50.028 100.667 70.428 71.000 10.125 30.208 100.530 90.406 70.337 70.000 90.000 60.585 30.742 70.500 40.000 60.000 30.472 21.000 10.417 90.563 40.631 70.275 30.000 80.800 60.841 10.000 40.083 10.000 50.174 80.000 20.055 60.667 20.000 60.000 70.250 61.000 10.286 70.058 80.391 70.209 60.313 40.167 10.278 110.200 80.083 50.000 50.200 80.264 70.000 40.250 70.714 10.500 10.196 70.333 60.500 90.750 10.668 10.500 10.000 20.500 90.333 81.000 10.000 10.000 80.438 10.500 10.000 31.000 10.333 70.226 20.250 50.250 50.000 30.000 30.668 50.000 10.174 100.000 10.000 80.750 10.000 10.667 80.000 10.000 10.638 80.333 60.579 20.000 60.333 30.000 11.000 10.000 10.063 80.385 70.600 10.647 20.066 80.264 60.469 50.246 30.000 70.000 10.264 40.000 10.000 20.000 61.000 10.125 10.000 50.000 20.200 40.000 50.000 41.000 10.000 2
Mask3D Scannet2000.388 60.542 60.357 70.237 70.808 50.676 50.741 70.832 70.496 50.151 90.628 70.021 50.955 40.578 60.753 40.612 60.591 60.822 110.609 70.926 40.614 80.291 60.725 80.163 70.890 60.380 110.615 30.517 70.130 90.806 10.857 30.024 80.511 70.412 110.226 10.597 70.756 61.000 10.111 40.792 20.736 70.091 50.610 60.527 40.323 101.000 10.504 50.063 81.000 10.853 60.010 60.974 40.839 30.667 60.301 70.883 10.266 40.039 30.640 20.311 50.739 60.463 61.000 10.000 60.287 60.715 70.313 80.600 21.000 10.027 40.076 100.502 110.500 40.409 40.000 30.194 50.125 70.500 30.491 70.748 40.050 100.042 70.776 80.352 70.008 30.000 50.033 10.254 50.000 20.005 80.552 40.008 30.020 60.750 10.500 50.409 60.065 70.511 50.107 70.178 80.000 31.000 10.400 20.016 70.000 50.400 30.571 30.000 40.060 80.044 80.000 60.514 40.278 71.000 10.258 70.017 90.125 110.000 20.792 70.399 71.000 10.000 10.013 70.265 20.018 80.000 31.000 10.335 60.381 10.500 10.250 50.004 20.000 30.727 40.000 10.497 60.000 10.188 60.677 60.000 10.708 60.000 10.000 10.945 10.391 50.123 90.000 60.028 60.000 11.000 10.000 10.099 40.451 50.400 40.668 10.573 20.606 10.077 110.003 100.004 60.000 10.042 90.000 10.000 21.000 11.000 10.000 30.042 30.000 20.200 40.302 10.000 41.000 10.000 2
Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang, Bastian Leibe: Mask3D for 3D Semantic Instance Segmentation. ICRA 2023