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


Method Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
OccuSeg+Semantic0.764 10.758 110.796 30.839 20.746 10.907 10.562 10.850 20.680 10.672 10.978 10.610 10.335 10.777 10.819 40.847 10.830 10.691 10.972 10.885 10.727 1
MinkowskiNetpermissive0.736 20.859 20.818 20.832 30.709 30.840 30.521 30.853 10.660 20.643 20.951 40.544 30.286 80.731 20.893 10.675 60.772 50.683 20.874 110.852 30.727 1
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
SparseConvNet0.725 30.647 200.821 10.846 10.721 20.869 20.533 20.754 80.603 50.614 30.955 20.572 20.325 20.710 30.870 20.724 30.823 20.628 40.934 20.865 20.683 4
CU-Hybrid Net0.693 40.596 230.789 40.803 40.677 40.800 100.469 80.846 30.554 110.591 50.948 100.500 40.316 30.609 60.847 30.732 20.808 30.593 60.894 50.839 40.652 5
KP-FCNN0.684 50.847 40.758 90.784 70.647 60.814 70.473 60.772 60.605 40.594 40.935 190.450 100.181 230.587 80.805 60.690 50.785 40.614 50.882 70.819 50.632 8
H. Thomas, C. Qi, J. Deschaud, B. Marcotegui, F. Goulette, L. Guibas.: KPConv: Flexible and Deformable Convolution for Point Clouds. ICCV 2019
PointASNL0.666 60.703 160.781 50.751 120.655 50.830 40.471 70.769 70.474 200.537 80.951 40.475 60.279 90.635 50.698 90.675 60.751 80.553 140.816 180.806 80.703 3
PointConvpermissive0.666 60.781 80.759 80.699 130.644 70.822 60.475 50.779 50.564 100.504 140.953 30.428 150.203 180.586 90.754 70.661 90.753 70.588 70.902 30.813 70.642 6
Wenxuan Wu, Zhongang Qi, Li Fuxin: PointConv: Deep Convolutional Networks on 3D Point Clouds. CVPR 2019
DMC-Net0.653 80.771 90.701 130.801 50.619 90.807 90.463 90.680 150.495 170.520 110.940 150.452 90.301 40.496 170.816 50.664 80.719 90.563 110.822 160.799 100.638 7
MVPNetpermissive0.641 90.831 50.715 120.671 160.590 120.781 120.394 170.679 170.642 30.553 70.937 180.462 80.256 100.649 40.406 220.626 130.691 120.666 30.877 80.792 130.608 11
Maximilian Jaritz, Jiayuan Gu, Hao Su: Multi-view PointNet for 3D Scene Understanding. GMDL Workshop, ICCV 2019
SConv0.636 100.830 60.697 160.752 110.572 150.780 130.445 110.716 110.529 120.530 90.951 40.446 110.170 240.507 150.666 110.636 120.682 130.541 170.886 60.799 100.594 15
joint point-basedpermissive0.634 110.614 220.778 60.667 180.633 80.825 50.420 130.804 40.467 220.561 60.951 40.494 50.291 50.566 110.458 190.579 180.764 60.559 130.838 140.814 60.598 14
Hung-Yueh Chiang, Yen-Liang Lin, Yueh-Cheng Liu, Winston H. Hsu: A Unified Point-Based Framework for 3D Segmentation. 3DV 2019
MCCNNpermissive0.633 120.866 10.731 110.771 80.576 140.809 80.410 150.684 140.497 160.491 150.949 80.466 70.105 300.581 100.646 120.620 140.680 140.542 160.817 170.795 120.618 9
P. Hermosilla, T. Ritschel, P.P. Vazquez, A. Vinacua, T. Ropinski: Monte Carlo Convolution for Learning on Non-Uniformly Sampled Point Clouds. SIGGRAPH Asia 2018
3DSM_DMMF0.631 130.626 210.745 100.801 50.607 100.751 180.506 40.729 100.565 90.491 150.866 320.434 120.197 200.595 70.630 130.709 40.705 100.560 120.875 100.740 220.491 25
SIConv0.625 140.830 60.694 170.757 100.563 160.772 150.448 100.647 200.520 130.509 120.949 80.431 140.191 210.496 170.614 140.647 100.672 150.535 180.876 90.783 140.571 17
HPEIN0.618 150.729 130.668 200.647 190.597 110.766 160.414 140.680 150.520 130.525 100.946 110.432 130.215 150.493 190.599 150.638 110.617 230.570 90.897 40.806 80.605 12
Li Jiang, Hengshuang Zhao, Shu Liu, Xiaoyong Shen, Chi-Wing Fu, Jiaya Jia: Hierarchical Point-Edge Interaction Network for Point Cloud Semantic Segmentation. ICCV 2019
SPH3D-GCNpermissive0.610 160.858 30.772 70.489 290.532 170.792 110.404 160.643 210.570 80.507 130.935 190.414 160.046 340.510 140.702 80.602 150.705 100.549 150.859 130.773 160.534 21
Huan Lei, Naveed Akhtar, and Ajmal Mian: Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds.
LAP-D0.594 170.720 140.692 180.637 210.456 230.773 140.391 190.730 90.587 60.445 200.940 150.381 180.288 60.434 210.453 200.591 160.649 170.581 80.777 220.749 210.610 10
DPC0.592 180.720 140.700 140.602 240.480 200.762 170.380 210.713 120.585 70.437 210.940 150.369 200.288 60.434 210.509 180.590 170.639 210.567 100.772 230.755 190.592 16
Francis Engelmann, Theodora Kontogianni, Bastian Leibe: Dilated Point Convolutions. arXiv
CCRFNet0.589 190.766 100.659 230.683 150.470 220.740 200.387 200.620 220.490 180.476 170.922 240.355 230.245 110.511 130.511 170.571 190.643 190.493 220.872 120.762 170.600 13
TextureNetpermissive0.566 200.672 180.664 210.671 160.494 180.719 210.445 110.678 180.411 270.396 220.935 190.356 220.225 130.412 230.535 160.565 200.636 220.464 240.794 210.680 270.568 18
Jingwei Huang, Haotian Zhang, Li Yi, Thomas Funkerhouser, Matthias Niessner, Leonidas Guibas: TextureNet: Consistent Local Parametrizations for Learning from High-Resolution Signals on Meshes. CVPR
DVVNet0.562 210.648 190.700 140.770 90.586 130.687 250.333 230.650 190.514 150.475 180.906 290.359 210.223 140.340 260.442 210.422 290.668 160.501 200.708 270.779 150.534 21
Pointnet++ & Featurepermissive0.557 220.735 120.661 220.686 140.491 190.744 190.392 180.539 250.451 230.375 240.946 110.376 190.205 170.403 240.356 240.553 210.643 190.497 210.824 150.756 180.515 23
PanopticFusion-label0.529 230.491 300.688 190.604 230.386 260.632 300.225 340.705 130.434 250.293 280.815 330.348 240.241 120.499 160.669 100.507 220.649 170.442 280.796 200.602 320.561 19
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
3DMV, FTSDF0.501 240.558 270.608 280.424 330.478 210.690 240.246 300.586 230.468 210.450 190.911 270.394 170.160 260.438 200.212 300.432 280.541 280.475 230.742 250.727 230.477 26
PCNN0.498 250.559 260.644 250.560 270.420 250.711 230.229 320.414 260.436 240.352 250.941 140.324 250.155 270.238 300.387 230.493 230.529 290.509 190.813 190.751 200.504 24
3DMV0.484 260.484 310.538 310.643 200.424 240.606 330.310 240.574 240.433 260.378 230.796 340.301 260.214 160.537 120.208 310.472 270.507 320.413 310.693 280.602 320.539 20
Angela Dai, Matthias Niessner: 3DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentation. ECCV'18
PointCNN with RGBpermissive0.458 270.577 250.611 270.356 350.321 320.715 220.299 260.376 290.328 320.319 260.944 130.285 280.164 250.216 330.229 290.484 250.545 270.456 260.755 240.709 240.475 27
Yangyan Li, Rui Bu, Mingchao Sun, Baoquan Chen: PointCNN. NeurIPS 2018
FCPNpermissive0.447 280.679 170.604 290.578 260.380 270.682 260.291 270.106 350.483 190.258 330.920 250.258 300.025 350.231 320.325 250.480 260.560 260.463 250.725 260.666 290.231 35
Dario Rethage, Johanna Wald, Jürgen Sturm, Nassir Navab, Federico Tombari: Fully-Convolutional Point Networks for Large-Scale Point Clouds. ECCV 2018
SurfaceConvPF0.442 290.505 290.622 260.380 340.342 310.654 280.227 330.397 280.367 300.276 300.924 230.240 310.198 190.359 250.262 270.366 310.581 240.435 290.640 300.668 280.398 29
Hao Pan, Shilin Liu, Yang Liu, Xin Tong: Convolutional Neural Networks on 3D Surfaces Using Parallel Frames.
PNET20.442 290.548 280.548 300.597 250.363 290.628 310.300 250.292 300.374 290.307 270.881 310.268 290.186 220.238 300.204 320.407 300.506 330.449 270.667 290.620 310.462 28
Tangent Convolutionspermissive0.438 310.437 330.646 240.474 300.369 280.645 290.353 220.258 320.282 340.279 290.918 260.298 270.147 280.283 270.294 260.487 240.562 250.427 300.619 310.633 300.352 31
Maxim Tatarchenko, Jaesik Park, Vladlen Koltun, Qian-Yi Zhou: Tangent convolutions for dense prediction in 3d. CVPR 2018
SPLAT Netcopyleft0.393 320.472 320.511 320.606 220.311 330.656 270.245 310.405 270.328 320.197 340.927 220.227 330.000 370.001 370.249 280.271 360.510 300.383 330.593 320.699 250.267 33
Hang Su, Varun Jampani, Deqing Sun, Subhransu Maji, Evangelos Kalogerakis, Ming-Hsuan Yang, Jan Kautz: SPLATNet: Sparse Lattice Networks for Point Cloud Processing. CVPR 2018
ScanNet+FTSDF0.383 330.297 350.491 330.432 320.358 300.612 320.274 280.116 340.411 270.265 310.904 300.229 320.079 320.250 280.185 330.320 340.510 300.385 320.548 330.597 340.394 30
PointNet++permissive0.339 340.584 240.478 340.458 310.256 350.360 360.250 290.247 330.278 350.261 320.677 360.183 340.117 290.212 340.145 350.364 320.346 360.232 360.548 330.523 350.252 34
Charles R. Qi, Li Yi, Hao Su, Leonidas J. Guibas: pointnet++: deep hierarchical feature learning on point sets in a metric space.
SSC-UNetpermissive0.308 350.353 340.290 360.278 360.166 360.553 340.169 360.286 310.147 360.148 360.908 280.182 350.064 330.023 360.018 370.354 330.363 340.345 340.546 350.685 260.278 32
ScanNetpermissive0.306 360.203 360.366 350.501 280.311 330.524 350.211 350.002 370.342 310.189 350.786 350.145 360.102 310.245 290.152 340.318 350.348 350.300 350.460 360.437 360.182 36
Angela Dai, Angel X. Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, Matthias Nießner: ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. CVPR'17
ERROR0.054 370.000 370.041 370.172 370.030 370.062 370.001 370.035 360.004 370.051 370.143 370.019 370.003 360.041 350.050 360.003 370.054 370.018 370.005 370.264 370.082 37

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




Method Infoavg ap 50%bathtubbedbookshelfcabinetchaircountercurtaindeskdoorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwindow
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
OccuSeg+instance0.672 11.000 10.758 60.682 30.576 10.842 10.477 10.504 100.524 20.567 10.585 10.451 20.557 41.000 10.751 10.797 20.563 51.000 10.467 3
GICN0.638 21.000 10.895 10.800 10.480 50.676 50.144 20.737 20.354 50.447 30.400 80.365 50.700 11.000 10.569 30.836 10.599 21.000 10.473 2
PointGroup0.636 31.000 10.765 40.624 40.505 40.797 20.116 50.696 30.384 40.441 40.559 20.476 10.596 31.000 10.666 20.756 40.556 70.997 60.513 1
MPA0.611 41.000 10.833 20.765 20.526 20.756 40.136 40.588 70.470 30.438 50.432 60.358 60.650 20.857 60.429 60.765 30.557 61.000 10.430 5
SSEN0.575 51.000 10.761 50.473 90.477 60.795 30.066 80.529 80.658 10.460 20.461 30.380 40.331 110.859 50.401 90.692 60.653 11.000 10.348 9
MTML0.549 61.000 10.807 30.588 70.327 90.647 60.004 150.815 10.180 90.418 60.364 100.182 90.445 61.000 10.442 50.688 70.571 41.000 10.396 6
Jean Lahoud, Bernard Ghanem, Marc Pollefeys, Martin R. Oswald: 3D Instance Segmentation via Multi-task Metric Learning. ICCV 2019 [oral]
Occipital-SCS0.512 71.000 10.716 80.509 80.506 30.611 80.092 70.602 60.177 100.346 90.383 90.165 100.442 70.850 90.386 100.618 90.543 80.889 110.389 7
3D-BoNet0.488 81.000 10.672 110.590 60.301 100.484 140.098 60.620 40.306 60.341 100.259 120.125 120.434 80.796 100.402 80.499 140.513 90.909 100.439 4
Bo Yang, Jianan Wang, Ronald Clark, Qingyong Hu, Sen Wang, Andrew Markham, Niki Trigoni: Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds. NeurIPS 2019 Spotlight
PanopticFusion-inst0.478 90.667 110.712 100.595 50.259 120.550 120.000 180.613 50.175 110.250 130.434 40.437 30.411 100.857 60.485 40.591 120.267 160.944 80.359 8
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
ResNet-backbone0.459 101.000 10.737 70.159 170.259 110.587 100.138 30.475 110.217 80.416 70.408 70.128 110.315 120.714 110.411 70.536 130.590 30.873 130.304 10
MASCpermissive0.447 110.528 150.555 130.381 100.382 70.633 70.002 160.509 90.260 70.361 80.432 50.327 70.451 50.571 120.367 110.639 80.386 100.980 70.276 11
Chen Liu, Yasutaka Furukawa: MASC: Multi-scale Affinity with Sparse Convolution for 3D Instance Segmentation.
3D-SISpermissive0.382 121.000 10.432 150.245 130.190 130.577 110.013 130.263 130.033 170.320 110.240 130.075 140.422 90.857 60.117 150.699 50.271 150.883 120.235 13
Ji Hou, Angela Dai, Matthias Niessner: 3D-SIS: 3D Semantic Instance Segmentation of RGB-D Scans. CVPR 2019
UNet-backbone0.319 130.667 110.715 90.233 140.189 140.479 150.008 140.218 140.067 160.201 140.173 140.107 130.123 150.438 130.150 130.615 100.355 110.916 90.093 18
R-PointNet0.306 140.500 160.405 160.311 110.348 80.589 90.054 90.068 170.126 120.283 120.290 110.028 160.219 130.214 160.331 120.396 160.275 140.821 150.245 12
RegionNet0.248 150.667 110.437 140.188 150.153 150.491 130.000 180.208 150.094 140.153 150.099 160.057 150.217 140.119 170.039 180.466 150.302 130.640 160.140 15
3D-BEVIS0.248 150.667 110.566 120.076 180.035 190.394 160.027 110.035 180.098 130.099 170.030 180.025 170.098 160.375 140.126 140.604 110.181 170.854 140.171 14
Cathrin Elich, Francis Engelmann, Jonas Schult, Theodora Kontogianni, Bastian Leibe: 3D-BEVIS: Birds-Eye-View Instance Segmentation.
Seg-Clusterpermissive0.215 170.370 170.337 180.285 120.105 160.325 170.025 120.282 120.085 150.105 160.107 150.007 190.079 170.317 150.114 160.309 180.304 120.587 170.123 17
Sgpn_scannet0.143 180.208 190.390 170.169 160.065 170.275 180.029 100.069 160.000 180.087 180.043 170.014 180.027 190.000 180.112 170.351 170.168 180.438 180.138 16
MaskRCNN 2d->3d Proj0.058 190.333 180.002 190.000 190.053 180.002 190.002 170.021 190.000 180.045 190.024 190.238 80.065 180.000 180.014 190.107 190.020 190.110 190.006 19

This table lists the benchmark results for the 2D semantic label scenario.


Method Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
CU-Hybrid-2D Net0.636 10.825 10.820 10.179 130.648 10.463 10.549 10.742 10.676 10.628 10.961 10.420 10.379 30.684 10.381 70.732 10.723 10.599 10.827 50.851 10.634 1
DMMF_3d0.605 20.651 40.744 50.782 10.637 20.387 20.536 20.732 20.590 30.540 20.856 90.359 40.306 80.596 40.539 10.627 80.706 20.497 40.785 90.757 80.476 9
DMMF0.597 30.543 80.755 40.749 20.585 40.338 40.494 40.704 40.598 20.494 80.911 40.347 60.327 70.593 50.527 20.675 40.646 60.513 20.842 30.774 60.527 7
MCA-Net0.595 40.533 90.756 30.746 30.590 30.334 60.506 30.670 50.587 40.500 60.905 60.366 30.352 40.601 30.506 40.669 70.648 40.501 30.839 40.769 70.516 8
RFBNet0.592 50.616 50.758 20.659 40.581 50.330 70.469 50.655 80.543 70.524 30.924 20.355 50.336 60.572 60.479 50.671 50.648 40.480 50.814 70.814 20.614 3
DCRedNet0.583 60.682 30.723 60.542 70.510 80.310 90.451 60.668 60.549 60.520 40.920 30.375 20.446 10.528 80.417 60.670 60.577 100.478 60.862 20.806 30.628 2
SSMAcopyleft0.577 70.695 20.716 80.439 90.563 60.314 80.444 70.719 30.551 50.503 50.887 80.346 70.348 50.603 20.353 90.709 20.600 80.457 80.901 10.786 40.599 4
Abhinav Valada, Rohit Mohan, Wolfram Burgard: Self-Supervised Model Adaptation for Multimodal Semantic Segmentation. International Journal of Computer Vision, 2019
FuseNetpermissive0.535 80.570 70.681 90.182 120.512 70.290 100.431 80.659 70.504 90.495 70.903 70.308 80.428 20.523 90.365 80.676 30.621 70.470 70.762 100.779 50.541 6
Caner Hazirbas, Lingni Ma, Csaba Domokos, Daniel Cremers: FuseNet: Incorporating Depth into Semantic Segmentation via Fusion-based CNN Architecture. ACCV 2016
AdapNet++copyleft0.503 90.613 60.722 70.418 100.358 130.337 50.370 110.479 110.443 100.368 110.907 50.207 110.213 120.464 110.525 30.618 90.657 30.450 90.788 80.721 100.408 12
Abhinav Valada, Rohit Mohan, Wolfram Burgard: Self-Supervised Model Adaptation for Multimodal Semantic Segmentation. International Journal of Computer Vision, 2019
3DMV (2d proj)0.498 100.481 110.612 100.579 60.456 100.343 30.384 90.623 90.525 80.381 100.845 100.254 100.264 100.557 70.182 110.581 110.598 90.429 100.760 110.661 120.446 11
Angela Dai, Matthias Niessner: 3DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentation. ECCV'18
ILC-PSPNet0.475 110.490 100.581 110.289 110.507 90.067 130.379 100.610 100.417 120.435 90.822 120.278 90.267 90.503 100.228 100.616 100.533 110.375 110.820 60.729 90.560 5
Enet (reimpl)0.376 120.264 130.452 130.452 80.365 110.181 110.143 130.456 120.409 130.346 120.769 130.164 120.218 110.359 120.123 130.403 130.381 130.313 130.571 120.685 110.472 10
Re-implementation of Adam Paszke, Abhishek Chaurasia, Sangpil Kim, Eugenio Culurciello: ENet: A Deep Neural Network Architecture for Real-Time Semantic Segmentation.
ScanNet (2d proj)permissive0.330 130.293 120.521 120.657 50.361 120.161 120.250 120.004 130.440 110.183 130.836 110.125 130.060 130.319 130.132 120.417 120.412 120.344 120.541 130.427 130.109 13
Angela Dai, Angel X. Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, Matthias Nießner: ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. CVPR'17

This table lists the benchmark results for the 2D semantic instance scenario.




Method Infoavg apbathtubbedbookshelfcabinetchaircountercurtaindeskdoorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwindow
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
MaskRCNN_ScanNetpermissive0.119 10.129 10.212 10.002 10.112 10.148 10.014 10.205 10.044 10.066 10.078 10.095 10.142 10.030 10.128 10.139 10.080 10.459 10.057 1
Re-implementation of Kaiming He, Georgia Gkioxari, Piotr Dollár, Ross Girshick: Mask R-CNN. ICCV'17

This table lists the benchmark results for the scene type classification scenario.




Method Infoavg recallapartmentbathroombedroom / hotelbookstore / libraryconference roomcopy/mail roomhallwaykitchenlaundry roomliving room / loungemiscofficestorage / basement / garage
sorted bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort by
MTL0.700 10.500 11.000 10.882 20.500 11.000 11.000 10.500 11.000 11.000 10.778 10.000 20.938 10.000 1
SE-ResNeXt-SSMA0.498 20.000 30.812 20.941 10.500 10.500 20.500 20.500 10.429 30.500 20.667 20.500 10.625 20.000 1
Abhinav Valada, Rohit Mohan, Wolfram Burgard: Self-Supervised Model Adaptation for Multimodal Semantic Segmentation. arXiv
resnet50_scannet0.353 30.250 20.812 20.529 30.500 10.500 20.000 30.500 10.571 20.000 30.556 30.000 20.375 30.000 1