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 aphead apcommon aptail apchairtabledoorcouchcabinetshelfdeskoffice 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
CompetitorFormer-2000.328 40.439 30.303 40.223 40.771 10.456 50.663 30.673 40.259 60.182 40.455 20.373 10.722 20.504 30.450 50.774 30.469 10.945 10.380 40.820 10.479 30.312 10.641 40.143 50.786 30.346 40.356 40.534 50.120 60.658 30.655 10.049 40.464 20.428 70.014 30.465 70.650 10.850 50.076 60.083 60.808 20.044 60.543 40.271 60.712 31.000 10.454 30.183 20.831 30.730 50.010 30.471 50.575 40.421 30.390 20.663 50.192 50.047 10.820 10.243 50.441 50.303 21.000 10.000 50.277 40.620 30.427 30.312 40.000 80.011 40.123 50.569 50.430 30.562 10.000 20.353 20.083 50.500 20.358 50.396 50.120 50.082 50.868 10.518 30.000 30.004 30.001 40.137 60.000 20.019 30.366 20.000 50.083 30.500 20.444 50.119 70.099 10.110 60.400 10.178 50.000 20.689 30.400 10.125 40.065 30.314 60.384 10.044 20.256 30.484 40.333 10.345 10.243 20.632 30.487 50.013 70.333 10.000 21.000 10.472 30.835 30.000 10.116 20.000 40.500 10.000 30.069 50.237 40.000 50.500 10.267 40.000 30.050 20.452 60.000 10.475 20.000 10.677 20.400 50.000 10.555 50.000 10.000 10.679 20.060 80.171 51.000 10.103 40.000 10.667 10.000 10.088 20.296 50.305 50.444 30.221 40.208 30.192 60.069 30.140 10.000 10.043 60.000 10.043 10.111 40.556 10.000 30.054 10.000 20.322 20.025 30.000 41.000 10.000 2
Volt-SPFormerpermissive0.367 20.475 20.359 20.248 20.678 30.494 30.736 10.689 30.416 30.170 50.484 10.008 50.663 30.575 10.524 10.787 10.418 20.928 20.550 10.684 50.470 40.308 20.685 10.193 30.799 10.565 10.365 30.560 30.144 40.682 20.556 40.052 30.663 10.417 80.000 40.527 40.609 21.000 10.299 10.000 70.831 10.051 50.635 20.524 10.650 41.000 10.442 40.235 10.873 20.817 20.004 50.383 80.693 20.469 20.348 30.682 40.380 20.012 50.400 60.240 60.664 20.284 31.000 10.125 20.329 30.660 20.717 10.318 20.250 30.029 20.340 20.748 10.333 50.407 20.000 20.017 60.556 21.000 10.552 10.549 20.238 20.099 40.821 30.515 40.000 30.000 40.014 10.232 30.111 10.013 50.333 40.002 40.000 60.139 70.389 60.822 10.029 50.551 10.247 40.230 40.000 20.719 20.378 40.500 20.778 10.400 30.117 40.000 30.388 10.439 50.278 20.192 50.241 30.537 50.588 40.466 30.333 10.000 21.000 10.395 41.000 10.000 10.013 40.000 40.254 40.000 30.556 20.710 10.000 50.500 10.304 30.000 30.000 30.864 10.000 10.502 10.000 10.500 30.588 20.000 10.655 30.000 10.000 10.652 30.764 20.112 70.250 30.278 30.000 10.222 30.000 10.050 50.528 20.533 10.345 50.638 10.167 70.066 90.117 20.019 30.000 10.113 40.000 10.000 20.444 10.556 10.000 30.028 20.000 20.156 40.000 40.167 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-Scannet200copyleft0.346 30.437 40.353 30.229 30.729 20.536 20.659 40.733 10.431 10.264 20.388 40.001 70.764 10.529 20.462 40.669 40.411 30.925 30.371 60.766 30.545 10.263 40.574 50.257 10.714 50.504 30.325 50.726 10.206 20.618 40.628 20.066 20.297 40.558 20.000 40.732 10.594 40.940 30.199 20.558 30.752 30.174 10.687 10.470 20.921 10.764 80.345 60.142 40.731 70.780 30.138 10.514 30.712 10.556 10.417 10.719 20.407 10.042 20.292 100.456 10.245 90.266 41.000 10.042 40.247 50.446 40.373 40.241 50.049 60.000 50.328 40.536 60.417 40.000 50.000 20.764 10.000 70.500 20.406 30.520 30.045 70.442 20.803 40.681 10.000 30.000 40.000 50.251 20.000 20.027 20.083 70.000 50.303 10.306 30.889 20.551 30.094 20.264 30.361 20.253 30.000 20.611 40.400 10.516 10.000 40.599 20.279 20.000 30.346 20.642 10.111 40.282 30.183 40.664 20.750 10.378 50.333 10.500 10.514 60.593 10.708 40.000 10.238 10.000 40.250 50.111 10.000 70.484 20.000 50.250 40.585 10.000 30.063 10.487 50.000 10.365 40.000 10.772 10.639 10.000 10.769 20.000 10.000 10.545 50.655 30.000 90.250 30.014 60.000 10.222 30.000 10.082 30.618 10.156 70.384 40.436 30.130 80.246 50.049 60.009 40.000 10.192 30.000 10.000 20.000 50.477 50.028 20.000 40.000 20.156 40.000 40.000 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
ACGP-ScanNet2000.381 10.486 10.362 10.275 10.643 60.543 10.676 20.647 80.365 50.284 10.435 30.031 30.649 40.449 40.514 20.782 20.400 40.895 40.480 20.772 20.423 70.291 30.678 20.242 20.753 40.524 20.412 10.694 20.209 10.612 60.444 70.080 10.329 30.395 90.000 40.538 30.608 31.000 10.062 70.903 10.733 40.133 20.597 30.388 50.795 21.000 10.466 20.179 30.926 10.824 10.007 40.494 40.652 30.391 40.330 40.779 10.114 60.032 30.497 40.307 40.752 10.314 11.000 10.394 10.346 20.673 10.462 20.313 30.778 10.077 10.454 10.635 30.486 10.170 40.001 10.069 50.556 20.500 20.546 20.686 10.541 10.461 10.821 20.548 20.037 10.009 20.000 50.301 10.000 20.018 40.304 50.007 30.197 20.248 50.792 30.581 20.008 80.380 20.288 30.336 10.000 20.731 10.400 10.500 20.194 20.600 10.112 60.500 10.240 50.512 30.000 50.312 20.247 10.569 40.677 20.574 10.333 10.000 20.792 40.486 21.000 10.000 10.037 30.000 40.396 30.000 30.556 20.438 30.036 30.146 70.396 20.000 30.000 30.832 20.000 10.406 30.000 10.365 40.499 30.000 10.815 10.000 10.000 10.785 10.850 10.143 61.000 10.472 10.000 10.139 60.000 10.104 10.512 30.349 40.483 20.184 50.197 60.500 10.065 50.060 20.000 10.271 10.000 10.000 20.444 10.556 10.000 30.000 40.000 20.252 30.144 20.014 31.000 10.002 1
TD3D Scannet200permissive0.211 70.332 70.177 70.103 70.662 40.413 60.463 70.705 20.192 80.145 60.266 60.215 20.452 90.209 70.222 100.219 100.315 60.893 50.380 50.617 60.439 50.047 90.646 30.080 70.610 70.253 50.237 70.293 70.135 50.379 100.494 50.048 50.252 60.451 40.184 20.483 50.395 60.852 40.083 50.551 40.278 70.036 70.337 70.266 70.544 60.963 40.079 100.039 50.740 60.604 70.000 70.586 10.283 70.282 70.059 70.633 70.028 70.004 60.559 30.309 30.420 60.028 101.000 10.000 50.456 10.411 50.372 50.060 90.046 70.000 50.040 90.694 20.083 70.000 50.000 20.000 70.000 70.083 90.252 70.260 90.200 40.160 30.669 60.111 70.000 30.000 40.006 30.169 50.000 20.007 60.296 60.032 10.074 40.139 70.000 70.321 50.031 40.108 70.088 70.157 60.000 20.231 90.026 100.000 70.000 40.356 50.052 70.000 30.240 60.147 60.000 50.015 70.046 80.144 80.073 80.414 40.222 90.000 20.806 30.343 70.486 70.000 10.008 50.038 30.083 60.002 20.028 60.074 60.032 40.150 60.039 70.008 10.000 30.250 90.000 10.125 80.000 10.052 60.260 80.000 10.143 100.000 10.000 10.543 60.207 60.404 10.000 50.003 70.000 10.000 70.000 10.037 60.093 90.272 60.342 60.039 90.281 20.249 40.224 10.000 60.000 10.074 50.000 10.000 20.000 50.278 70.000 30.000 40.889 10.323 10.000 40.014 20.000 70.000 2
Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: Top-Down Beats Bottom-Up in 3D Instance Segmentation. WACV 2024
Minkowski 34D Inst.permissive0.130 90.246 90.083 90.043 100.547 90.236 90.415 90.672 50.141 100.133 80.067 90.000 80.521 70.114 100.238 90.289 70.232 90.883 60.182 100.373 100.486 20.076 80.488 90.022 90.529 80.199 90.110 90.217 90.100 70.460 90.319 90.000 80.025 100.472 30.000 40.394 80.210 90.537 90.004 90.000 70.083 100.000 100.299 90.061 100.201 100.761 90.084 90.008 80.720 80.557 100.000 70.317 100.280 80.094 100.020 100.564 100.000 90.000 70.400 60.048 80.259 80.101 81.000 10.000 50.190 80.142 100.094 100.137 80.089 50.000 50.101 60.355 100.000 80.000 50.000 20.000 70.000 70.444 60.082 100.384 60.000 100.000 70.334 100.004 100.000 30.000 40.000 50.041 90.000 20.000 80.026 100.000 50.000 60.000 90.000 70.082 90.022 70.000 100.021 90.088 90.000 20.241 80.033 90.000 70.000 40.067 80.000 100.000 30.000 80.000 80.000 50.000 90.026 90.262 70.016 90.000 90.278 50.000 20.500 80.394 50.028 100.000 10.000 70.000 40.000 80.000 30.000 70.019 90.000 50.000 80.000 80.000 30.000 30.156 100.000 10.032 100.000 10.000 70.194 100.000 10.248 90.000 10.000 10.099 90.019 90.308 30.000 50.000 80.000 10.000 70.000 10.007 90.122 70.000 80.175 80.063 60.000 90.271 20.000 100.000 60.000 10.000 100.000 10.000 20.000 50.278 70.000 30.000 40.000 20.111 70.000 40.000 40.000 70.000 2
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
LGround Inst.permissive0.154 80.275 80.108 80.060 80.573 70.381 70.434 80.654 70.190 90.141 70.097 80.000 80.503 80.180 80.252 80.242 90.242 80.881 70.448 30.494 80.429 60.078 70.364 100.024 80.654 60.213 80.222 80.239 80.099 80.616 50.363 80.000 80.092 80.444 50.000 40.383 90.209 100.815 60.030 80.000 70.166 80.002 90.295 100.099 90.364 70.778 50.177 80.001 90.427 100.585 90.000 70.470 60.268 100.205 80.045 80.642 60.007 80.000 70.333 90.148 70.407 70.130 71.000 10.000 50.156 90.189 80.097 90.169 60.000 80.000 50.056 70.400 80.000 80.000 50.000 20.000 70.556 20.278 80.203 80.323 80.019 90.000 70.402 90.026 80.000 30.000 40.000 50.044 80.000 20.000 80.037 90.000 50.000 60.181 60.000 70.127 60.006 90.028 90.023 80.115 70.000 20.327 60.267 60.000 70.000 40.000 90.028 80.000 30.000 80.000 80.000 50.003 80.048 70.135 90.222 70.089 60.278 50.000 20.514 60.333 80.611 60.000 10.000 70.000 40.000 80.000 30.000 70.037 80.000 50.000 80.000 80.000 30.000 30.322 70.000 10.209 60.000 10.000 70.278 70.000 10.302 80.000 10.000 10.143 80.148 70.000 90.000 50.000 80.000 10.000 70.000 10.015 80.064 100.000 80.272 70.031 100.000 90.257 30.028 70.000 60.000 10.041 70.000 10.000 20.000 50.222 100.000 30.000 40.000 20.000 100.000 40.000 40.000 70.000 2
David Rozenberszki, Or Litany, Angela Dai: Language-Grounded Indoor 3D Semantic Segmentation in the Wild.
ODIN - Ins200permissive0.265 60.349 60.268 50.163 60.485 100.366 80.549 60.492 100.421 20.229 30.265 70.003 60.609 60.297 60.320 60.327 60.251 70.848 80.314 90.526 70.324 90.138 60.529 60.178 40.440 90.186 100.306 60.546 40.160 30.494 80.476 60.016 60.231 70.594 10.000 40.615 20.357 70.630 80.141 30.167 50.665 50.054 40.360 60.451 40.610 50.769 70.640 10.032 60.746 50.698 60.040 20.389 70.550 60.371 50.257 60.617 80.310 30.000 70.481 50.022 90.463 30.160 61.000 10.125 20.193 70.267 70.253 70.156 70.000 80.000 50.332 30.606 40.444 20.000 50.000 20.281 31.000 10.417 70.344 60.238 100.218 30.000 70.655 70.506 50.000 30.052 10.000 50.091 70.000 20.035 10.370 10.000 50.000 60.250 40.903 10.037 100.031 30.221 40.197 50.285 20.037 10.191 100.200 70.083 50.000 40.200 70.115 50.000 30.250 40.552 20.278 20.077 60.107 60.389 60.674 30.565 20.278 50.000 20.361 100.333 80.361 80.000 10.000 70.438 10.451 20.000 31.000 10.074 60.204 20.250 40.250 50.000 30.000 30.493 40.000 10.083 90.000 10.000 70.317 60.000 10.481 60.000 10.000 10.188 70.333 50.345 20.000 50.333 20.000 10.333 20.000 10.035 70.266 60.478 20.506 10.054 70.205 40.119 80.067 40.000 60.000 10.210 20.000 10.000 20.000 50.389 60.097 10.000 40.000 20.111 70.000 40.000 40.889 60.000 2
CSC-Pretrain Inst.permissive0.123 100.223 100.082 100.046 90.564 80.152 100.394 100.578 90.235 70.116 100.034 100.000 80.348 100.119 90.297 70.285 80.202 100.838 90.323 80.407 90.184 100.037 100.516 70.013 100.424 100.214 70.093 100.105 100.078 100.542 70.250 100.000 80.064 90.444 50.000 40.224 100.231 80.537 90.001 100.000 70.126 90.004 80.308 80.193 80.244 90.343 100.228 70.000 100.441 90.588 80.000 70.338 90.275 90.189 90.030 90.600 90.000 90.000 70.378 80.000 100.108 100.098 91.000 10.000 50.096 100.172 90.144 80.011 100.125 40.000 50.000 100.376 90.000 80.000 50.000 20.000 70.000 70.042 100.141 90.377 70.051 60.000 70.483 80.017 90.000 30.000 40.000 50.022 100.000 20.000 80.065 80.000 50.000 60.000 90.000 70.094 80.000 100.042 80.000 100.064 100.000 20.259 70.089 80.000 70.000 40.000 90.022 90.000 30.000 80.000 80.000 50.000 90.018 100.111 100.000 100.000 90.278 50.000 20.444 90.333 80.333 90.000 10.000 70.000 40.000 80.000 30.000 70.000 100.000 50.000 80.000 80.000 30.000 30.267 80.000 10.184 70.000 10.000 70.211 90.000 10.378 70.000 10.000 10.063 100.000 100.275 40.000 50.000 80.000 10.000 70.000 10.007 100.105 80.000 80.032 100.045 80.198 50.171 70.028 70.000 60.000 10.006 80.000 10.000 20.000 50.278 70.000 30.000 40.000 20.044 90.000 40.000 40.000 70.000 2
Ji Hou, Benjamin Graham, Matthias Nießner, Saining Xie: Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts. CVPR 2021
Mask3D Scannet2000.278 50.383 50.263 60.168 50.661 50.465 40.572 50.665 60.391 40.121 90.304 50.015 40.647 50.349 50.474 30.489 50.321 50.816 100.351 70.722 40.402 80.195 50.515 80.082 60.795 20.215 60.396 20.377 60.082 90.724 10.586 30.015 70.277 50.377 100.201 10.475 60.572 50.778 70.089 40.759 20.556 60.068 30.506 50.467 30.323 80.778 50.427 50.027 70.789 40.744 40.003 60.570 20.561 50.337 60.265 50.711 30.258 40.031 40.569 20.311 20.441 40.179 51.000 10.000 50.233 60.411 60.283 60.380 10.667 20.016 30.048 80.418 70.139 60.173 30.000 20.086 40.014 60.500 20.384 40.497 40.044 80.032 60.752 50.287 60.003 20.000 40.007 20.208 40.000 20.001 70.349 30.008 20.014 50.509 10.500 40.323 40.023 60.176 50.107 60.105 80.000 20.605 50.378 40.016 60.000 40.400 30.192 30.000 30.048 70.037 70.000 50.275 40.119 50.810 10.258 60.006 80.083 100.000 20.568 50.377 60.708 40.000 10.005 60.147 20.014 70.000 30.556 20.085 50.325 10.500 10.083 60.004 20.000 30.590 30.000 10.365 50.000 10.116 50.491 40.000 10.626 40.000 10.000 10.579 40.391 40.050 80.000 50.028 50.000 10.222 30.000 10.063 40.302 40.356 30.149 90.573 20.415 10.013 100.002 90.004 50.000 10.005 90.000 10.000 20.444 10.514 40.000 30.028 20.000 20.156 40.267 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