3D Semantic Label Results - ScanNet Benchmark

The 3D semantic labeling task involves predicting a semantic labeling of a 3D scan mesh.

Evaluation and metrics

Our evaluation ranks all methods according to the PASCAL VOC intersection-over-union metric (IoU). IoU = TP/(TP+FP+FN), where TP, FP, and FN are the numbers of true positive, false positive, and false negative pixels, respectively. Predicted labels are evaluated per-vertex over the respective 3D scan mesh; for 3D approaches that operate on other representations like grids or points, the predicted labels should be mapped onto the mesh vertices (e.g., one such example for grid to mesh vertices is provided in the evaluation helpers).

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

Method	avg iou	bathtub	bed	bookshelf	cabinet	chair	counter	curtain	desk	door	floor	otherfurniture	picture	refrigerator	shower curtain	sink	sofa	table	toilet	wall	window

OccuSeg+Semantic	0.764 1	0.758 31	0.796 7	0.839 4	0.746 2	0.907 1	0.562 1	0.850 5	0.680 2	0.672 1	0.978 1	0.610 1	0.335 3	0.777 1	0.819 16	0.847 1	0.830 1	0.691 3	0.972 1	0.885 1	0.727 2

BPNet	0.749 2	0.909 1	0.818 4	0.811 8	0.752 1	0.839 6	0.485 14	0.842 7	0.673 3	0.644 4	0.957 3	0.528 8	0.305 9	0.773 2	0.859 4	0.788 2	0.818 3	0.693 2	0.916 6	0.856 5	0.723 4
Wenbo Hu, Hengshuang Zhao, Li Jiang, Jiaya Jia, Tien-Tsin Wong: Bidirectional Projection Network for Cross Dimension Scene Understanding. CVPR 2021 (Oral)
Virtual MVFusion	0.746 3	0.771 26	0.819 3	0.848 2	0.702 8	0.865 3	0.397 45	0.899 1	0.699 1	0.664 2	0.948 20	0.588 2	0.330 4	0.746 5	0.851 8	0.764 3	0.796 7	0.704 1	0.935 4	0.866 2	0.728 1
Abhijit Kundu, Xiaoqi Yin, Alireza Fathi, David Ross, Brian Brewington, Thomas Funkhouser, Caroline Pantofaru: Virtual Multi-view Fusion for 3D Semantic Segmentation. ECCV 2020
VMNet	0.746 3	0.870 6	0.838 1	0.858 1	0.729 4	0.850 4	0.501 8	0.874 2	0.587 21	0.658 3	0.956 4	0.564 4	0.299 10	0.765 3	0.900 1	0.716 8	0.812 4	0.631 11	0.939 2	0.858 4	0.709 5
Zeyu HU, Xuyang Bai, Jiaxiang Shang, Runze Zhang, Jiayu Dong, Xin Wang, Guangyuan Sun, Hongbo Fu, Chiew-Lan Tai: VMNet: Voxel-Mesh Network for Geodesic-Aware 3D Semantic Segmentation. ICCV 2021 (Oral)
MinkowskiNet	0.736 5	0.859 8	0.818 4	0.832 5	0.709 7	0.840 5	0.521 5	0.853 4	0.660 4	0.643 5	0.951 11	0.544 5	0.286 15	0.731 6	0.893 2	0.675 19	0.772 14	0.683 4	0.874 31	0.852 6	0.727 2
C. Choy, J. Gwak, S. Savarese: 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. CVPR 2019
SparseConvNet	0.725 6	0.647 50	0.821 2	0.846 3	0.721 5	0.869 2	0.533 3	0.754 23	0.603 18	0.614 6	0.955 5	0.572 3	0.325 5	0.710 7	0.870 3	0.724 7	0.823 2	0.628 12	0.934 5	0.865 3	0.683 8

MatchingNet	0.724 7	0.812 18	0.812 6	0.810 9	0.735 3	0.834 8	0.495 11	0.860 3	0.572 27	0.602 12	0.954 6	0.512 10	0.280 16	0.757 4	0.845 11	0.725 6	0.780 11	0.606 20	0.937 3	0.851 7	0.700 7

RFCR	0.702 8	0.889 3	0.745 23	0.813 7	0.672 12	0.818 19	0.493 12	0.815 10	0.623 10	0.610 8	0.947 24	0.470 20	0.249 28	0.594 20	0.848 9	0.705 13	0.779 12	0.646 7	0.892 18	0.823 15	0.611 24
Jingyu Gong, Jiachen Xu, Xin Tan, Haichuan Song, Yanyun Qu, Yuan Xie, Lizhuang Ma: Omni-Supervised Point Cloud Segmentation via Gradual Receptive Field Component Reasoning. CVPR2021
One Thing One Click	0.701 9	0.825 16	0.796 7	0.723 29	0.716 6	0.832 9	0.433 35	0.816 8	0.634 8	0.609 9	0.969 2	0.418 43	0.344 2	0.559 31	0.833 12	0.715 9	0.808 5	0.560 34	0.902 10	0.847 8	0.680 9

JSENet	0.699 10	0.881 5	0.762 15	0.821 6	0.667 13	0.800 31	0.522 4	0.792 15	0.613 11	0.607 10	0.935 46	0.492 14	0.205 39	0.576 26	0.853 6	0.691 14	0.758 19	0.652 6	0.872 34	0.828 12	0.649 15
Zeyu HU, Mingmin Zhen, Xuyang BAI, Hongbo Fu, Chiew-lan Tai: JSENet: Joint Semantic Segmentation and Edge Detection Network for 3D Point Clouds. ECCV 2020
Feature-Geometry Net	0.694 11	0.894 2	0.741 26	0.768 21	0.677 10	0.827 13	0.491 13	0.811 11	0.612 12	0.612 7	0.948 20	0.464 23	0.250 26	0.554 32	0.828 13	0.708 12	0.781 10	0.614 15	0.884 22	0.822 17	0.593 35
Kangcheng Liu, Ben M. Chen: https://arxiv.org/abs/2012.09439. arXiv Preprint
CU-Hybrid Net	0.693 12	0.596 56	0.789 9	0.803 11	0.677 10	0.800 31	0.469 18	0.846 6	0.554 35	0.591 15	0.948 20	0.500 12	0.316 7	0.609 17	0.847 10	0.732 4	0.808 5	0.593 25	0.894 16	0.839 9	0.652 14

Feature_GeometricNet	0.690 13	0.884 4	0.754 20	0.795 14	0.647 18	0.818 19	0.422 36	0.802 14	0.612 12	0.604 11	0.945 29	0.462 24	0.189 45	0.563 30	0.853 6	0.726 5	0.765 16	0.632 10	0.904 8	0.821 18	0.606 28
Kangcheng Liu, Ben M. Chen: https://arxiv.org/abs/2012.09439. arXiv Preprint
FusionNet	0.688 14	0.704 43	0.741 26	0.754 26	0.656 14	0.829 11	0.501 8	0.741 28	0.609 15	0.548 22	0.950 15	0.522 9	0.371 1	0.633 15	0.756 22	0.715 9	0.771 15	0.623 13	0.861 39	0.814 20	0.658 12
Feihu Zhang, Jin Fang, Benjamin Wah, Philip Torr: Deep FusionNet for Point Cloud Semantic Segmentation. ECCV 2020
VACNN++	0.684 15	0.728 39	0.757 19	0.776 18	0.690 9	0.804 29	0.464 22	0.816 8	0.577 25	0.587 16	0.945 29	0.508 11	0.276 18	0.671 8	0.710 30	0.663 24	0.750 22	0.589 26	0.881 26	0.832 11	0.653 13

KP-FCNN	0.684 15	0.847 11	0.758 18	0.784 16	0.647 18	0.814 22	0.473 16	0.772 18	0.605 16	0.594 14	0.935 46	0.450 31	0.181 48	0.587 21	0.805 18	0.690 15	0.785 9	0.614 15	0.882 24	0.819 19	0.632 19
H. Thomas, C. Qi, J. Deschaud, B. Marcotegui, F. Goulette, L. Guibas.: KPConv: Flexible and Deformable Convolution for Point Clouds. ICCV 2019
Superpoint Network	0.683 17	0.851 10	0.728 33	0.800 13	0.653 16	0.806 28	0.468 19	0.804 12	0.572 27	0.602 12	0.946 26	0.453 30	0.239 31	0.519 40	0.822 14	0.689 17	0.762 18	0.595 24	0.895 15	0.827 13	0.630 20

PointContrast_LA_SEM	0.683 17	0.757 32	0.784 10	0.786 15	0.639 21	0.824 16	0.408 40	0.775 17	0.604 17	0.541 24	0.934 50	0.532 6	0.269 21	0.552 33	0.777 20	0.645 32	0.793 8	0.640 8	0.913 7	0.824 14	0.671 10

VI-PointConv	0.676 19	0.770 28	0.754 20	0.783 17	0.621 24	0.814 22	0.552 2	0.758 21	0.571 29	0.557 20	0.954 6	0.529 7	0.268 23	0.530 38	0.682 35	0.675 19	0.719 28	0.603 21	0.888 20	0.833 10	0.665 11
Xingyi Li, Wenxuan Wu, Xiaoli Z. Fern, Li Fuxin: The Devils in the Point Clouds: Studying the Robustness of Point Cloud Convolutions.
ROSMRF3D	0.673 20	0.789 19	0.748 22	0.763 24	0.635 22	0.814 22	0.407 42	0.747 25	0.581 24	0.573 17	0.950 15	0.484 15	0.271 20	0.607 18	0.754 23	0.649 28	0.774 13	0.596 23	0.883 23	0.823 15	0.606 28

SALANet	0.670 21	0.816 17	0.770 14	0.768 21	0.652 17	0.807 27	0.451 24	0.747 25	0.659 5	0.545 23	0.924 53	0.473 19	0.149 59	0.571 28	0.811 17	0.635 35	0.746 23	0.623 13	0.892 18	0.794 31	0.570 41

PointConv	0.666 22	0.781 21	0.759 17	0.699 34	0.644 20	0.822 17	0.475 15	0.779 16	0.564 32	0.504 36	0.953 8	0.428 38	0.203 41	0.586 23	0.754 23	0.661 25	0.753 20	0.588 27	0.902 10	0.813 22	0.642 16
Wenxuan Wu, Zhongang Qi, Li Fuxin: PointConv: Deep Convolutional Networks on 3D Point Clouds. CVPR 2019
PointASNL	0.666 22	0.703 44	0.781 11	0.751 28	0.655 15	0.830 10	0.471 17	0.769 19	0.474 50	0.537 25	0.951 11	0.475 18	0.279 17	0.635 13	0.698 34	0.675 19	0.751 21	0.553 39	0.816 49	0.806 23	0.703 6
Xu Yan, Chaoda Zheng, Zhen Li, Sheng Wang, Shuguang Cui: PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Sampling. CVPR 2020
DCM-Net	0.658 24	0.778 22	0.702 38	0.806 10	0.619 25	0.813 25	0.468 19	0.693 40	0.494 46	0.524 30	0.941 38	0.449 32	0.298 11	0.510 42	0.821 15	0.675 19	0.727 27	0.568 31	0.826 46	0.803 25	0.637 18
Jonas Schult, Francis Engelmann, Theodora Kontogianni, Bastian Leibe: DualConvMesh-Net: Joint Geodesic and Euclidean Convolutions on 3D Meshes. CVPR 2020 [Oral]
HPGCNN	0.656 25	0.698 45	0.743 25	0.650 48	0.564 40	0.820 18	0.505 7	0.758 21	0.631 9	0.479 42	0.945 29	0.480 16	0.226 32	0.572 27	0.774 21	0.690 15	0.735 26	0.614 15	0.853 42	0.776 43	0.597 33
Jisheng Dang, Qingyong Hu, Yulan Guo, Jun Yang: HPGCNN.
SAFNet-seg	0.654 26	0.752 33	0.734 29	0.664 46	0.583 34	0.815 21	0.399 44	0.754 23	0.639 7	0.535 26	0.942 36	0.470 20	0.309 8	0.665 9	0.539 44	0.650 27	0.708 33	0.635 9	0.857 41	0.793 32	0.642 16
Linqing Zhao, Jiwen Lu, Jie Zhou: Similarity-Aware Fusion Network for 3D Semantic Segmentation. IROS 2021
RandLA-Net	0.645 27	0.778 22	0.731 30	0.699 34	0.577 35	0.829 11	0.446 28	0.736 29	0.477 49	0.523 32	0.945 29	0.454 28	0.269 21	0.484 49	0.749 26	0.618 39	0.738 24	0.599 22	0.827 45	0.792 34	0.621 22

MVPNet	0.641 28	0.831 13	0.715 34	0.671 43	0.590 31	0.781 41	0.394 46	0.679 43	0.642 6	0.553 21	0.937 43	0.462 24	0.256 24	0.649 10	0.406 56	0.626 36	0.691 40	0.666 5	0.877 27	0.792 34	0.608 27
Maximilian Jaritz, Jiayuan Gu, Hao Su: Multi-view PointNet for 3D Scene Understanding. GMDL Workshop, ICCV 2019
PointConv-SFPN	0.641 28	0.776 24	0.703 37	0.721 30	0.557 43	0.826 14	0.451 24	0.672 45	0.563 33	0.483 41	0.943 35	0.425 41	0.162 54	0.644 11	0.726 27	0.659 26	0.709 32	0.572 29	0.875 29	0.786 38	0.559 46

PointMRNet	0.640 30	0.717 42	0.701 39	0.692 36	0.576 36	0.801 30	0.467 21	0.716 34	0.563 33	0.459 46	0.953 8	0.429 37	0.169 51	0.581 24	0.854 5	0.605 40	0.710 31	0.550 40	0.894 16	0.793 32	0.575 39

FPConv	0.639 31	0.785 20	0.760 16	0.713 33	0.603 28	0.798 34	0.392 47	0.534 58	0.603 18	0.524 30	0.948 20	0.457 26	0.250 26	0.538 36	0.723 28	0.598 43	0.696 39	0.614 15	0.872 34	0.799 26	0.567 43
Yiqun Lin, Zizheng Yan, Haibin Huang, Dong Du, Ligang Liu, Shuguang Cui, Xiaoguang Han: FPConv: Learning Local Flattening for Point Convolution. CVPR 2020
PointSPNet	0.637 32	0.734 36	0.692 44	0.714 32	0.576 36	0.797 35	0.446 28	0.743 27	0.598 20	0.437 50	0.942 36	0.403 45	0.150 58	0.626 16	0.800 19	0.649 28	0.697 38	0.557 37	0.846 43	0.777 42	0.563 44

SConv	0.636 33	0.830 14	0.697 42	0.752 27	0.572 39	0.780 42	0.445 30	0.716 34	0.529 39	0.530 27	0.951 11	0.446 33	0.170 50	0.507 44	0.666 37	0.636 34	0.682 42	0.541 45	0.886 21	0.799 26	0.594 34

Supervoxel-CNN	0.635 34	0.656 48	0.711 35	0.719 31	0.613 26	0.757 48	0.444 32	0.765 20	0.534 38	0.566 18	0.928 51	0.478 17	0.272 19	0.636 12	0.531 46	0.664 23	0.645 50	0.508 49	0.864 38	0.792 34	0.611 24

joint point-based	0.634 35	0.614 53	0.778 12	0.667 45	0.633 23	0.825 15	0.420 37	0.804 12	0.467 52	0.561 19	0.951 11	0.494 13	0.291 12	0.566 29	0.458 50	0.579 48	0.764 17	0.559 36	0.838 44	0.814 20	0.598 32
Hung-Yueh Chiang, Yen-Liang Lin, Yueh-Cheng Liu, Winston H. Hsu: A Unified Point-Based Framework for 3D Segmentation. 3DV 2019
MCCNN	0.633 36	0.866 7	0.731 30	0.771 19	0.576 36	0.809 26	0.410 39	0.684 41	0.497 45	0.491 38	0.949 17	0.466 22	0.105 64	0.581 24	0.646 38	0.620 37	0.680 43	0.542 44	0.817 48	0.795 29	0.618 23
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
PointMTL	0.632 37	0.731 37	0.688 47	0.675 40	0.591 30	0.784 40	0.444 32	0.565 56	0.610 14	0.492 37	0.949 17	0.456 27	0.254 25	0.587 21	0.706 31	0.599 42	0.665 47	0.612 19	0.868 37	0.791 37	0.579 38

APCF-Net	0.631 38	0.742 34	0.687 49	0.672 41	0.557 43	0.792 37	0.408 40	0.665 46	0.545 36	0.508 34	0.952 10	0.428 38	0.186 46	0.634 14	0.702 32	0.620 37	0.706 34	0.555 38	0.873 33	0.798 28	0.581 37
Haojia, Lin: Adaptive Pyramid Context Fusion for Point Cloud Perception. GRSL
PointNet2-SFPN	0.631 38	0.771 26	0.692 44	0.672 41	0.524 46	0.837 7	0.440 34	0.706 38	0.538 37	0.446 48	0.944 33	0.421 42	0.219 35	0.552 33	0.751 25	0.591 45	0.737 25	0.543 43	0.901 12	0.768 46	0.557 47

3DSM_DMMF	0.631 38	0.626 52	0.745 23	0.801 12	0.607 27	0.751 49	0.506 6	0.729 32	0.565 31	0.491 38	0.866 64	0.434 34	0.197 43	0.595 19	0.630 39	0.709 11	0.705 35	0.560 34	0.875 29	0.740 53	0.491 56

FusionAwareConv	0.630 41	0.604 55	0.741 26	0.766 23	0.590 31	0.747 50	0.501 8	0.734 30	0.503 44	0.527 28	0.919 57	0.454 28	0.323 6	0.550 35	0.420 55	0.678 18	0.688 41	0.544 42	0.896 14	0.795 29	0.627 21
Jiazhao Zhang, Chenyang Zhu, Lintao Zheng, Kai Xu: Fusion-Aware Point Convolution for Online Semantic 3D Scene Segmentation. CVPR 2020
SIConv	0.625 42	0.830 14	0.694 43	0.757 25	0.563 41	0.772 44	0.448 26	0.647 49	0.520 41	0.509 33	0.949 17	0.431 36	0.191 44	0.496 47	0.614 40	0.647 31	0.672 45	0.535 47	0.876 28	0.783 39	0.571 40

HPEIN	0.618 43	0.729 38	0.668 50	0.647 50	0.597 29	0.766 45	0.414 38	0.680 42	0.520 41	0.525 29	0.946 26	0.432 35	0.215 36	0.493 48	0.599 41	0.638 33	0.617 55	0.570 30	0.897 13	0.806 23	0.605 30
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-GCN	0.610 44	0.858 9	0.772 13	0.489 63	0.532 45	0.792 37	0.404 43	0.643 50	0.570 30	0.507 35	0.935 46	0.414 44	0.046 69	0.510 42	0.702 32	0.602 41	0.705 35	0.549 41	0.859 40	0.773 44	0.534 49
Huan Lei, Naveed Akhtar, and Ajmal Mian: Spherical Kernel for Efficient Graph Convolution on 3D Point Clouds. TPAMI 2020
AttAN	0.609 45	0.760 30	0.667 51	0.649 49	0.521 47	0.793 36	0.457 23	0.648 48	0.528 40	0.434 52	0.947 24	0.401 46	0.153 57	0.454 50	0.721 29	0.648 30	0.717 29	0.536 46	0.904 8	0.765 47	0.485 57
Gege Zhang, Qinghua Ma, Licheng Jiao, Fang Liu and Qigong Sun: AttAN: Attention Adversarial Networks for 3D Point Cloud Semantic Segmentation. IJCAI2020
LAP-D	0.594 46	0.720 40	0.692 44	0.637 52	0.456 54	0.773 43	0.391 49	0.730 31	0.587 21	0.445 49	0.940 40	0.381 49	0.288 13	0.434 53	0.453 51	0.591 45	0.649 48	0.581 28	0.777 53	0.749 52	0.610 26

DPC	0.592 47	0.720 40	0.700 40	0.602 55	0.480 51	0.762 47	0.380 52	0.713 36	0.585 23	0.437 50	0.940 40	0.369 51	0.288 13	0.434 53	0.509 48	0.590 47	0.639 53	0.567 32	0.772 54	0.755 50	0.592 36
Francis Engelmann, Theodora Kontogianni, Bastian Leibe: Dilated Point Convolutions: On the Receptive Field Size of Point Convolutions on 3D Point Clouds. ICRA 2020
SegGCN	0.589 48	0.833 12	0.731 30	0.539 60	0.514 48	0.789 39	0.448 26	0.467 60	0.573 26	0.484 40	0.936 44	0.396 47	0.061 68	0.501 45	0.507 49	0.594 44	0.700 37	0.563 33	0.874 31	0.771 45	0.493 55
Huan Lei, Naveed Akhtar, and Ajmal Mian: SegGCN: Efficient 3D Point Cloud Segmentation with Fuzzy Spherical Kernel. CVPR 2020
CCRFNet	0.589 48	0.766 29	0.659 54	0.683 38	0.470 53	0.740 52	0.387 51	0.620 52	0.490 47	0.476 43	0.922 55	0.355 55	0.245 29	0.511 41	0.511 47	0.571 49	0.643 51	0.493 52	0.872 34	0.762 48	0.600 31

ROSMRF	0.580 50	0.772 25	0.707 36	0.681 39	0.563 41	0.764 46	0.362 53	0.515 59	0.465 53	0.465 45	0.936 44	0.427 40	0.207 38	0.438 51	0.577 42	0.536 53	0.675 44	0.486 53	0.723 58	0.779 40	0.524 51

TextureNet	0.566 51	0.672 47	0.664 52	0.671 43	0.494 49	0.719 53	0.445 30	0.678 44	0.411 60	0.396 55	0.935 46	0.356 54	0.225 33	0.412 55	0.535 45	0.565 50	0.636 54	0.464 56	0.794 52	0.680 60	0.568 42
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
DVVNet	0.562 52	0.648 49	0.700 40	0.770 20	0.586 33	0.687 57	0.333 56	0.650 47	0.514 43	0.475 44	0.906 61	0.359 53	0.223 34	0.340 59	0.442 54	0.422 62	0.668 46	0.501 50	0.708 59	0.779 40	0.534 49

Pointnet++ & Feature	0.557 53	0.735 35	0.661 53	0.686 37	0.491 50	0.744 51	0.392 47	0.539 57	0.451 55	0.375 58	0.946 26	0.376 50	0.205 39	0.403 56	0.356 59	0.553 52	0.643 51	0.497 51	0.824 47	0.756 49	0.515 52

PointMRNet-lite	0.553 54	0.633 51	0.648 56	0.659 47	0.430 56	0.800 31	0.390 50	0.592 53	0.454 54	0.371 59	0.939 42	0.368 52	0.136 61	0.368 57	0.448 52	0.560 51	0.715 30	0.486 53	0.882 24	0.720 55	0.462 60

PanopticFusion-label	0.529 55	0.491 64	0.688 47	0.604 54	0.386 59	0.632 62	0.225 67	0.705 39	0.434 58	0.293 63	0.815 65	0.348 56	0.241 30	0.499 46	0.669 36	0.507 55	0.649 48	0.442 61	0.796 51	0.602 66	0.561 45
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
Online SegFusion	0.515 56	0.607 54	0.644 59	0.579 57	0.434 55	0.630 63	0.353 54	0.628 51	0.440 56	0.410 53	0.762 68	0.307 58	0.167 52	0.520 39	0.403 57	0.516 54	0.565 58	0.447 60	0.678 61	0.701 57	0.514 53
Davide Menini, Suryansh Kumar, Martin R. Oswald, Erik Sandstroem, Cristian Sminchisescu, Luc van Gool: A Real-Time Learning Framework for Joint 3D Reconstruction and Semantic Segmentation. Robotics and Automation Letters Submission
3DMV, FTSDF	0.501 57	0.558 60	0.608 63	0.424 68	0.478 52	0.690 56	0.246 63	0.586 54	0.468 51	0.450 47	0.911 59	0.394 48	0.160 55	0.438 51	0.212 65	0.432 61	0.541 63	0.475 55	0.742 56	0.727 54	0.477 58

PCNN	0.498 58	0.559 59	0.644 59	0.560 59	0.420 58	0.711 55	0.229 65	0.414 61	0.436 57	0.352 60	0.941 38	0.324 57	0.155 56	0.238 65	0.387 58	0.493 56	0.529 64	0.509 48	0.813 50	0.751 51	0.504 54

3DMV	0.484 59	0.484 65	0.538 66	0.643 51	0.424 57	0.606 66	0.310 57	0.574 55	0.433 59	0.378 57	0.796 66	0.301 59	0.214 37	0.537 37	0.208 66	0.472 60	0.507 67	0.413 64	0.693 60	0.602 66	0.539 48
Angela Dai, Matthias Niessner: 3DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentation. ECCV'18
PointCNN with RGB	0.458 60	0.577 58	0.611 62	0.356 70	0.321 66	0.715 54	0.299 59	0.376 64	0.328 67	0.319 61	0.944 33	0.285 61	0.164 53	0.216 68	0.229 64	0.484 58	0.545 62	0.456 58	0.755 55	0.709 56	0.475 59
Yangyan Li, Rui Bu, Mingchao Sun, Baoquan Chen: PointCNN. NeurIPS 2018
FCPN	0.447 61	0.679 46	0.604 64	0.578 58	0.380 60	0.682 58	0.291 60	0.106 70	0.483 48	0.258 68	0.920 56	0.258 64	0.025 70	0.231 67	0.325 60	0.480 59	0.560 60	0.463 57	0.725 57	0.666 62	0.231 70
Dario Rethage, Johanna Wald, Jürgen Sturm, Nassir Navab, Federico Tombari: Fully-Convolutional Point Networks for Large-Scale Point Clouds. ECCV 2018
PNET2	0.442 62	0.548 61	0.548 65	0.597 56	0.363 62	0.628 64	0.300 58	0.292 65	0.374 63	0.307 62	0.881 63	0.268 63	0.186 46	0.238 65	0.204 67	0.407 63	0.506 68	0.449 59	0.667 62	0.620 64	0.462 60

SurfaceConvPF	0.442 62	0.505 63	0.622 61	0.380 69	0.342 64	0.654 60	0.227 66	0.397 63	0.367 64	0.276 65	0.924 53	0.240 66	0.198 42	0.359 58	0.262 62	0.366 64	0.581 57	0.435 62	0.640 63	0.668 61	0.398 62
Hao Pan, Shilin Liu, Yang Liu, Xin Tong: Convolutional Neural Networks on 3D Surfaces Using Parallel Frames.
Tangent Convolutions	0.438 64	0.437 68	0.646 58	0.474 65	0.369 61	0.645 61	0.353 54	0.258 67	0.282 69	0.279 64	0.918 58	0.298 60	0.147 60	0.283 62	0.294 61	0.487 57	0.562 59	0.427 63	0.619 64	0.633 63	0.352 65
Maxim Tatarchenko, Jaesik Park, Vladlen Koltun, Qian-Yi Zhou: Tangent convolutions for dense prediction in 3d. CVPR 2018
3DWSSS	0.425 65	0.525 62	0.647 57	0.522 61	0.324 65	0.488 70	0.077 71	0.712 37	0.353 65	0.401 54	0.636 70	0.281 62	0.176 49	0.340 59	0.565 43	0.175 71	0.551 61	0.398 65	0.370 71	0.602 66	0.361 64

subcloud_weak	0.411 66	0.479 66	0.650 55	0.475 64	0.285 69	0.519 69	0.087 70	0.725 33	0.396 62	0.386 56	0.621 71	0.250 65	0.117 62	0.338 61	0.443 53	0.188 70	0.594 56	0.369 68	0.377 70	0.616 65	0.306 66

SPLAT Net	0.393 67	0.472 67	0.511 67	0.606 53	0.311 67	0.656 59	0.245 64	0.405 62	0.328 67	0.197 69	0.927 52	0.227 68	0.000 72	0.001 72	0.249 63	0.271 69	0.510 65	0.383 67	0.593 65	0.699 58	0.267 68
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+FTSDF	0.383 68	0.297 70	0.491 68	0.432 67	0.358 63	0.612 65	0.274 61	0.116 69	0.411 60	0.265 66	0.904 62	0.229 67	0.079 66	0.250 63	0.185 68	0.320 67	0.510 65	0.385 66	0.548 66	0.597 69	0.394 63

PointNet++	0.339 69	0.584 57	0.478 69	0.458 66	0.256 70	0.360 71	0.250 62	0.247 68	0.278 70	0.261 67	0.677 69	0.183 69	0.117 62	0.212 69	0.145 70	0.364 65	0.346 71	0.232 71	0.548 66	0.523 70	0.252 69
Charles R. Qi, Li Yi, Hao Su, Leonidas J. Guibas: pointnet++: deep hierarchical feature learning on point sets in a metric space.
SSC-UNet	0.308 70	0.353 69	0.290 71	0.278 71	0.166 71	0.553 67	0.169 69	0.286 66	0.147 71	0.148 71	0.908 60	0.182 70	0.064 67	0.023 71	0.018 72	0.354 66	0.363 69	0.345 69	0.546 68	0.685 59	0.278 67

ScanNet	0.306 71	0.203 71	0.366 70	0.501 62	0.311 67	0.524 68	0.211 68	0.002 72	0.342 66	0.189 70	0.786 67	0.145 71	0.102 65	0.245 64	0.152 69	0.318 68	0.348 70	0.300 70	0.460 69	0.437 71	0.182 71
Angela Dai, Angel X. Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, Matthias Nießner: ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. CVPR'17
ERROR	0.054 72	0.000 72	0.041 72	0.172 72	0.030 72	0.062 72	0.001 72	0.035 71	0.004 72	0.051 72	0.143 72	0.019 72	0.003 71	0.041 70	0.050 71	0.003 72	0.054 72	0.018 72	0.005 72	0.264 72	0.082 72

3D Semantic label benchmark