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 apalarm 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 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 bysort bysort bysort bysort bysort bysort by
Mask3D Scannet2000.278 10.383 10.263 10.168 10.506 10.068 10.083 40.000 10.000 10.000 10.023 10.149 30.302 10.778 20.647 10.569 10.500 10.031 10.014 20.027 10.173 10.311 10.195 10.351 20.258 10.000 10.082 10.000 10.003 10.037 10.391 11.000 10.000 10.014 10.000 10.572 10.573 10.661 10.000 10.003 10.005 30.082 30.349 10.028 10.000 10.605 10.515 20.509 10.711 11.000 10.665 20.015 10.107 10.402 30.201 10.083 10.304 10.759 10.491 10.378 10.572 10.119 10.277 10.013 40.089 10.283 10.411 10.267 10.006 20.156 10.000 10.116 10.000 10.105 20.556 10.514 10.396 10.275 10.323 10.215 10.380 10.000 10.000 10.356 10.005 10.208 10.325 10.000 10.050 30.400 10.561 10.258 10.179 10.722 10.147 10.000 10.586 10.063 10.015 10.139 10.016 10.028 10.708 10.418 10.016 10.048 30.500 10.489 10.349 10.001 10.475 10.086 10.365 10.000 10.500 10.000 10.323 20.000 10.222 10.000 10.497 10.626 10.044 20.795 10.556 10.008 10.121 30.265 10.667 10.789 10.568 10.579 10.444 10.176 10.004 10.474 10.752 10.233 10.014 10.002 30.570 10.007 10.377 40.000 10.000 10.000 10.337 10.000 10.000 10.384 10.465 10.287 10.085 10.048 10.816 40.467 10.810 10.377 10.415 10.744 10.000 10.004 10.724 10.778 10.590 10.000 10.032 10.441 10.000 10.377 20.391 10.427 10.321 10.192 1
Minkowski 34D Inst.permissive0.130 30.246 30.083 30.043 40.299 30.000 40.278 10.000 10.000 10.000 10.022 20.175 20.122 20.537 30.521 20.400 20.000 20.000 20.000 30.008 20.000 20.048 30.076 30.182 40.000 30.000 10.022 30.000 10.000 20.000 20.141 40.000 20.000 10.000 20.000 10.210 30.063 20.547 40.000 10.000 20.000 40.100 10.026 40.000 20.000 10.241 40.488 30.000 30.564 41.000 10.672 10.000 20.021 30.486 10.000 20.000 20.067 30.000 20.194 40.033 40.415 30.026 30.025 40.271 10.004 30.094 40.142 40.000 20.000 30.111 20.000 10.000 20.000 10.088 30.083 40.278 20.110 30.000 30.082 40.199 40.137 30.000 10.000 10.000 20.000 20.041 30.000 20.000 10.308 10.067 20.280 20.016 30.101 30.373 40.000 20.000 10.319 30.007 30.000 20.000 20.000 20.000 20.028 40.355 40.000 20.101 10.444 20.289 20.114 40.000 20.394 20.000 20.032 40.000 10.000 20.000 10.201 40.000 10.000 20.000 10.384 20.248 40.000 40.529 30.000 20.000 20.133 20.020 40.089 30.720 20.500 30.099 30.000 20.000 40.000 20.238 40.334 40.190 20.000 20.000 40.317 40.000 20.472 10.000 10.000 10.000 10.094 40.000 10.000 10.082 40.236 30.004 40.019 30.000 20.883 10.061 40.262 20.217 30.000 30.557 40.000 10.000 20.460 40.761 30.156 40.000 10.000 20.259 30.000 10.394 10.019 30.084 40.232 30.000 4
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
CSC-Pretrain Inst.permissive0.123 40.223 40.082 40.046 30.308 20.004 20.278 10.000 10.000 10.000 10.000 40.032 40.105 30.537 30.348 40.378 30.000 20.000 20.000 30.000 40.000 20.000 40.037 40.323 30.000 30.000 10.013 40.000 10.000 20.000 20.235 20.000 20.000 10.000 20.000 10.231 20.045 30.564 30.000 10.000 20.006 20.078 40.065 20.000 20.000 10.259 30.516 10.000 30.600 31.000 10.578 40.000 20.000 40.184 40.000 20.000 20.034 40.000 20.211 30.089 30.394 40.018 40.064 30.171 30.001 40.144 20.172 30.000 20.000 30.044 30.000 10.000 20.000 10.064 40.126 30.278 20.093 40.000 30.094 30.214 20.011 40.000 10.000 10.000 20.000 20.022 40.000 20.000 10.275 20.000 30.275 30.000 40.098 40.407 30.000 20.000 10.250 40.007 40.000 20.000 20.000 20.000 20.333 30.376 30.000 20.000 40.042 40.285 30.119 30.000 20.224 40.000 20.184 30.000 10.000 20.000 10.244 30.000 10.000 20.000 10.377 30.378 20.051 10.424 40.000 20.000 20.116 40.030 30.125 20.441 30.444 40.063 40.000 20.042 20.000 20.297 20.483 20.096 40.000 20.028 10.338 30.000 20.444 20.000 10.000 10.000 10.189 30.000 10.000 10.141 30.152 40.017 30.000 40.000 20.838 30.193 20.111 40.105 40.198 20.588 20.000 10.000 20.542 30.343 40.267 30.000 10.000 20.108 40.000 10.333 30.000 40.228 20.202 40.022 3
Ji Hou, Benjamin Graham, Matthias Nie├čner, Saining Xie: Exploring Data-Efficient 3D Scene Understanding with Contrastive Scene Contexts. CVPR 2021
LGround Inst.permissive0.154 20.275 20.108 20.060 20.295 40.002 30.278 10.000 10.000 10.000 10.006 30.272 10.064 40.815 10.503 30.333 40.000 20.000 20.556 10.001 30.000 20.148 20.078 20.448 10.007 20.000 10.024 20.000 10.000 20.000 20.190 30.000 20.000 10.000 20.000 10.209 40.031 40.573 20.000 10.000 20.041 10.099 20.037 30.000 20.000 10.327 20.364 40.181 20.642 21.000 10.654 30.000 20.023 20.429 20.000 20.000 20.097 20.000 20.278 20.267 20.434 20.048 20.092 20.257 20.030 20.097 30.189 20.000 20.089 10.000 40.000 10.000 20.000 10.115 10.166 20.222 40.222 20.003 20.127 20.213 30.169 20.000 10.000 10.000 20.000 20.044 20.000 20.000 10.000 40.000 30.268 40.222 20.130 20.494 20.000 20.000 10.363 20.015 20.000 20.000 20.000 20.000 20.611 20.400 20.000 20.056 20.278 30.242 40.180 20.000 20.383 30.000 20.209 20.000 10.000 20.000 10.364 10.000 10.000 20.000 10.323 40.302 30.019 30.654 20.000 20.000 20.141 10.045 20.000 40.427 40.514 20.143 20.000 20.028 30.000 20.252 30.402 30.156 30.000 20.028 10.470 20.000 20.444 20.000 10.000 10.000 10.205 20.000 10.000 10.203 20.381 20.026 20.037 20.000 20.881 20.099 30.135 30.239 20.000 30.585 30.000 10.000 20.616 20.778 10.322 20.000 10.000 20.407 20.000 10.333 30.148 20.177 30.242 20.028 2
David Rozenberszki, Or Litany, Angela Dai: Language-Grounded Indoor 3D Semantic Segmentation in the Wild.