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 with limited annotations scenario.




Method Infoavg ioubathtubbedbookshelfcabinetchaircountercurtaindeskdoorfloorotherfurniturepicturerefrigeratorshower curtainsinksofatabletoiletwallwindow
sort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysort bysorted bysort by
Q2E0.721 10.984 10.785 10.684 20.693 20.879 10.563 10.822 10.640 10.659 10.965 20.493 10.147 50.711 10.866 10.631 30.797 10.663 10.932 20.849 10.660 1
ActiveST0.703 20.977 20.776 30.657 50.707 10.874 20.541 20.744 20.605 20.610 20.968 10.442 40.126 60.705 20.785 20.742 10.791 20.586 20.940 10.839 20.645 2
Gengxin Liu, Oliver van Kaick, Hui Huang, Ruizhen Hu: Active Self-Training for Weakly Supervised 3D Scene Semantic Segmentation.
One-Thing-One-Click0.594 70.756 60.722 70.494 120.546 80.795 70.371 60.725 60.559 40.488 70.957 30.367 80.261 20.547 40.575 120.225 120.671 100.543 60.904 50.826 30.557 6
Zhengzhe Liu, Xiaojuan Qi, Chi-Wing Fu: One Thing One Click: A Self-Training Approach for Weakly Supervised 3D Semantic Segmentation. CVPR 2021
GaIA0.638 50.536 120.783 20.651 60.600 50.840 30.413 50.728 40.490 70.520 60.948 40.475 20.299 10.518 60.680 40.629 40.729 60.573 40.906 30.815 40.626 3
Min Seok Lee*, Seok Woo Yang*, and Sung Won Han: GaIA: Graphical Information gain based Attention Network for Weakly Supervised 3D Point Cloud Semantic Segmentation. WACV 2023
DE-3DLearner LA0.639 40.839 30.723 60.681 30.629 40.839 50.424 40.728 40.538 50.526 40.945 60.427 60.120 70.511 70.643 60.547 60.781 30.566 50.905 40.809 50.607 4
Ping-Chung Yu, Cheng Sun, Min Sun: Data Efficient 3D Learner via Knowledge Transferred from 2D Model. ECCV 2022
WS3D_LA_Sempermissive0.662 30.812 40.762 40.742 10.635 30.828 60.474 30.736 30.588 30.546 30.947 50.450 30.174 40.536 50.752 30.668 20.735 50.583 30.902 60.797 60.573 5
Kangcheng Liu: WS3D: Weakly Supervised 3D Scene Segmentation with Region-Level Boundary Awareness and Instance Discrimination. European Conference on Computer Vision (ECCV), 2022
LE0.608 60.791 50.726 50.651 60.589 60.779 90.346 90.662 80.493 60.524 50.923 130.430 50.234 30.572 30.638 70.411 100.708 70.533 80.855 70.782 70.508 7
VIBUSpermissive0.586 80.736 80.623 110.664 40.559 70.840 30.358 70.666 70.447 80.429 100.944 70.421 70.000 130.411 90.629 80.614 50.745 40.541 70.848 90.758 80.493 8
Beiwen Tian,Liyi Luo,Hao Zhao,Guyue Zhou: VIBUS: Data-efficient 3D Scene Parsing with VIewpoint Bottleneck and Uncertainty-Spectrum Modeling. ISPRS Journal of Photogrammetry and Remote Sensing
PointContrast_LA_SEM0.550 90.735 90.676 80.601 90.475 90.794 80.288 110.621 100.378 120.430 90.940 80.303 100.089 100.379 100.580 110.531 70.689 90.422 110.852 80.758 80.468 9
CSC_LA_SEM0.531 110.659 100.638 100.578 100.417 110.775 100.254 120.537 110.396 100.439 80.939 100.284 120.083 110.414 80.599 100.488 90.698 80.444 100.785 110.747 100.440 10
Viewpoint_BN_LA_AIR0.548 100.747 70.574 130.631 80.456 100.762 110.355 80.639 90.412 90.404 110.940 80.335 90.107 80.277 120.645 50.495 80.666 110.517 90.818 100.740 110.431 11
Liyi Luo, Beiwen Tian, Hao Zhao, Guyue Zhou: Pointly-supervised 3D Scene Parsing with Viewpoint Bottleneck.
Scratch_LA_SEM0.382 130.389 130.606 120.401 130.303 130.705 130.169 130.460 130.292 130.282 130.939 100.207 130.004 120.147 130.201 130.184 130.592 130.389 130.409 130.714 120.250 13
SQN_LA0.486 120.587 110.649 90.527 110.372 120.718 120.320 100.510 120.393 110.325 120.924 120.290 110.095 90.287 110.607 90.356 110.626 120.416 120.672 120.680 130.359 12