3D Semantic Instance Results - ScanNet Benchmark

The 3D semantic instance prediction task involves detecting and segmenting the object in an 3D scan mesh.

Evaluation and metrics

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). Note that multiple predictions of the same ground truth instance are penalized as false positives.

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

Method	avg ap 50%	bathtub	bed	bookshelf	cabinet	chair	counter	curtain	desk	door	otherfurniture	picture	refrigerator	shower curtain	sink	sofa	table	toilet	window

PointRel	0.816 1	1.000 1	0.971 7	0.908 6	0.743 2	0.923 6	0.573 6	0.714 22	0.695 17	0.734 8	0.747 2	0.725 10	0.809 1	1.000 1	0.814 8	0.899 3	0.820 3	1.000 1	0.610 16
: Relation3D (PointRel): Enhancing Relation Modeling for Point Cloud Instance Segmentation.
Spherical Mask(CtoF)	0.812 2	1.000 1	0.973 6	0.852 13	0.718 5	0.917 8	0.574 4	0.677 29	0.748 10	0.729 12	0.715 6	0.795 2	0.809 1	1.000 1	0.831 4	0.854 9	0.787 10	1.000 1	0.638 5

EV3D	0.811 3	1.000 1	0.968 8	0.852 13	0.717 6	0.921 7	0.574 5	0.677 29	0.748 10	0.730 11	0.703 11	0.795 2	0.809 1	1.000 1	0.831 4	0.854 9	0.778 14	1.000 1	0.638 6

SIM3D	0.803 4	1.000 1	0.967 9	0.863 12	0.692 17	0.924 5	0.552 10	0.732 21	0.667 21	0.732 10	0.662 15	0.796 1	0.789 9	1.000 1	0.803 9	0.864 6	0.766 19	1.000 1	0.643 4

OneFormer3D	0.801 5	1.000 1	0.973 5	0.909 5	0.698 14	0.928 3	0.582 3	0.668 34	0.685 18	0.780 2	0.687 13	0.698 18	0.702 14	1.000 1	0.794 11	0.900 2	0.784 12	0.986 51	0.635 7
Maxim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: OneFormer3D: One Transformer for Unified Point Cloud Segmentation.
Competitor-SPFormer	0.800 6	1.000 1	0.986 2	0.845 15	0.705 12	0.915 9	0.532 12	0.733 20	0.757 9	0.733 9	0.708 8	0.698 17	0.648 34	0.981 37	0.890 1	0.830 18	0.796 7	0.997 38	0.644 3

UniPerception	0.800 6	1.000 1	0.930 11	0.872 10	0.727 4	0.862 23	0.454 18	0.764 13	0.820 1	0.746 6	0.706 9	0.750 5	0.772 10	0.926 44	0.764 17	0.818 26	0.826 1	0.997 38	0.660 2

InsSSM	0.799 8	1.000 1	0.915 13	0.710 40	0.729 3	0.925 4	0.664 1	0.670 32	0.770 6	0.766 3	0.739 3	0.737 6	0.700 15	1.000 1	0.792 12	0.829 20	0.815 4	0.997 38	0.625 9
Lei Yao, Yi Wang, Moyun Liu, Lap-Pui Chau: SGIFormer: Semantic-guided and Geometric-enhanced Interleaving Transformer for 3D Instance Segmentation. TCSVT, 2024
TST3D	0.795 9	1.000 1	0.929 12	0.918 4	0.709 9	0.884 18	0.596 2	0.704 25	0.769 7	0.734 7	0.644 20	0.699 16	0.751 12	1.000 1	0.794 10	0.876 5	0.757 22	0.997 38	0.550 31
Duc Tran Dang Trung, Byeongkeun Kang, Yeejin Lee: MSTA3D: Multi-scale Twin-attention for 3D Instance Segmentation. ACM Multimedia 2024
MG-Former	0.791 10	1.000 1	0.980 4	0.837 18	0.626 25	0.897 11	0.543 11	0.759 15	0.800 5	0.766 4	0.659 16	0.769 4	0.697 18	1.000 1	0.791 13	0.707 47	0.791 9	1.000 1	0.610 15

ExtMask3D	0.789 11	1.000 1	0.988 1	0.756 33	0.706 11	0.912 10	0.429 19	0.647 39	0.806 4	0.755 5	0.673 14	0.689 19	0.772 11	1.000 1	0.789 14	0.852 11	0.811 5	1.000 1	0.617 12

Queryformer	0.787 12	1.000 1	0.933 10	0.601 49	0.754 1	0.886 16	0.558 9	0.661 36	0.767 8	0.665 18	0.716 5	0.639 24	0.808 5	1.000 1	0.844 3	0.897 4	0.804 6	1.000 1	0.624 10

MAFT	0.786 13	1.000 1	0.894 18	0.807 22	0.694 16	0.893 14	0.486 14	0.674 31	0.740 12	0.786 1	0.704 10	0.727 9	0.739 13	1.000 1	0.707 23	0.849 13	0.756 23	1.000 1	0.685 1

KmaxOneFormerNet	0.783 14	0.903 54	0.981 3	0.794 26	0.706 10	0.931 2	0.561 8	0.701 26	0.706 15	0.727 13	0.697 12	0.731 8	0.689 21	1.000 1	0.856 2	0.750 38	0.761 21	1.000 1	0.599 20

Mask3D	0.780 15	1.000 1	0.786 42	0.716 38	0.696 15	0.885 17	0.500 13	0.714 22	0.810 3	0.672 17	0.715 6	0.679 20	0.809 1	1.000 1	0.831 4	0.833 17	0.787 10	1.000 1	0.602 18
Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang, Bastian Leibe: Mask3D for 3D Semantic Instance Segmentation. ICRA 2023
SPFormer	0.770 16	0.903 54	0.903 15	0.806 23	0.609 31	0.886 15	0.568 7	0.815 6	0.705 16	0.711 14	0.655 17	0.652 23	0.685 22	1.000 1	0.789 15	0.809 27	0.776 16	1.000 1	0.583 24
Sun Jiahao, Qing Chunmei, Tan Junpeng, Xu Xiangmin: Superpoint Transformer for 3D Scene Instance Segmentation. AAAI 2023 [Oral]
SoftGroup++	0.769 17	1.000 1	0.803 35	0.937 1	0.684 18	0.865 20	0.213 34	0.870 2	0.664 22	0.571 24	0.758 1	0.702 14	0.807 6	1.000 1	0.653 30	0.902 1	0.792 8	1.000 1	0.626 8

SoftGroup	0.761 18	1.000 1	0.808 31	0.845 15	0.716 7	0.862 22	0.243 31	0.824 4	0.655 24	0.620 19	0.734 4	0.699 15	0.791 8	0.981 37	0.716 21	0.844 14	0.769 17	1.000 1	0.594 22
Thang Vu, Kookhoi Kim, Tung M. Luu, Xuan Thanh Nguyen, Chang D. Yoo: SoftGroup for 3D Instance Segmentaiton on Point Clouds. CVPR 2022 [Oral]
ISBNet	0.757 19	1.000 1	0.904 14	0.731 36	0.678 19	0.895 12	0.458 16	0.644 41	0.670 20	0.710 15	0.620 25	0.732 7	0.650 24	1.000 1	0.756 18	0.778 30	0.779 13	1.000 1	0.614 13
Tuan Duc Ngo, Binh-Son Hua, Khoi Nguyen: ISBNet: a 3D Point Cloud Instance Segmentation Network with Instance-aware Sampling and Box-aware Dynamic Convolution. CVPR 2023
TD3D	0.751 20	1.000 1	0.774 43	0.867 11	0.621 27	0.934 1	0.404 20	0.706 24	0.812 2	0.605 22	0.633 23	0.626 25	0.690 20	1.000 1	0.640 32	0.820 23	0.777 15	1.000 1	0.612 14
Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: Top-Down Beats Bottom-Up in 3D Instance Segmentation. WACV 2024
PBNet	0.747 21	1.000 1	0.818 27	0.837 19	0.713 8	0.844 25	0.457 17	0.647 39	0.711 14	0.614 20	0.617 27	0.657 22	0.650 24	1.000 1	0.692 24	0.822 22	0.765 20	1.000 1	0.595 21
W.Zhao, Y.Yan, C.Yang, J.Ye,X.Yang,K.Huang: Divide and Conquer: 3D Instance Segmentation With Point-Wise Binarization. ICCV 2023
GraphCut	0.732 22	1.000 1	0.788 40	0.724 37	0.642 24	0.859 24	0.248 30	0.787 11	0.618 27	0.596 23	0.653 19	0.722 12	0.583 46	1.000 1	0.766 16	0.861 7	0.825 2	1.000 1	0.504 37

IPCA-Inst	0.731 23	1.000 1	0.788 41	0.884 9	0.698 13	0.788 41	0.252 29	0.760 14	0.646 25	0.511 32	0.637 22	0.665 21	0.804 7	1.000 1	0.644 31	0.778 31	0.747 25	1.000 1	0.561 28

TopoSeg	0.725 24	1.000 1	0.806 34	0.933 2	0.668 21	0.758 45	0.272 28	0.734 19	0.630 26	0.549 28	0.654 18	0.606 26	0.697 19	0.966 41	0.612 36	0.839 15	0.754 24	1.000 1	0.573 25

DKNet	0.718 25	1.000 1	0.814 28	0.782 27	0.619 28	0.872 19	0.224 32	0.751 17	0.569 31	0.677 16	0.585 31	0.724 11	0.633 36	0.981 37	0.515 46	0.819 24	0.736 26	1.000 1	0.617 11
Yizheng Wu, Min Shi, Shuaiyuan Du, Hao Lu, Zhiguo Cao, Weicai Zhong: 3D Instances as 1D Kernels. ECCV 2022
SSEC	0.707 26	1.000 1	0.850 20	0.924 3	0.648 22	0.747 48	0.162 36	0.862 3	0.572 30	0.520 30	0.624 24	0.549 29	0.649 33	1.000 1	0.560 41	0.706 48	0.768 18	1.000 1	0.591 23

HAIS	0.699 27	1.000 1	0.849 21	0.820 20	0.675 20	0.808 35	0.279 26	0.757 16	0.465 37	0.517 31	0.596 29	0.559 28	0.600 40	1.000 1	0.654 29	0.767 33	0.676 30	0.994 47	0.560 29
Shaoyu Chen, Jiemin Fang, Qian Zhang, Wenyu Liu, Xinggang Wang: Hierarchical Aggregation for 3D Instance Segmentation. ICCV 2021
SSTNet	0.698 28	1.000 1	0.697 59	0.888 8	0.556 38	0.803 36	0.387 21	0.626 43	0.417 42	0.556 27	0.585 32	0.702 13	0.600 40	1.000 1	0.824 7	0.720 46	0.692 28	1.000 1	0.509 36
Zhihao Liang, Zhihao Li, Songcen Xu, Mingkui Tan, Kui Jia: Instance Segmentation in 3D Scenes using Semantic Superpoint Tree Networks. ICCV2021
DualGroup	0.694 29	1.000 1	0.799 37	0.811 21	0.622 26	0.817 30	0.376 22	0.805 9	0.590 29	0.487 36	0.568 35	0.525 33	0.650 24	0.835 54	0.600 37	0.829 19	0.655 33	1.000 1	0.526 33

SphereSeg	0.680 30	1.000 1	0.856 19	0.744 34	0.618 29	0.893 13	0.151 37	0.651 38	0.713 13	0.537 29	0.579 34	0.430 43	0.651 23	1.000 1	0.389 57	0.744 41	0.697 27	0.991 49	0.601 19

DANCENET	0.680 30	1.000 1	0.807 32	0.733 35	0.600 32	0.768 44	0.375 23	0.543 51	0.538 32	0.610 21	0.599 28	0.498 34	0.632 38	0.981 37	0.739 20	0.856 8	0.633 39	0.882 62	0.454 46

Box2Mask	0.677 32	1.000 1	0.847 22	0.771 29	0.509 47	0.816 31	0.277 27	0.558 50	0.482 34	0.562 26	0.640 21	0.448 39	0.700 15	1.000 1	0.666 25	0.852 12	0.578 46	0.997 38	0.488 41
Julian Chibane, Francis Engelmann, Tuan Anh Tran, Gerard Pons-Moll: Box2Mask: Weakly Supervised 3D Semantic Instance Segmentation Using Bounding Boxes. ECCV 2022
OccuSeg+instance	0.672 33	1.000 1	0.758 51	0.682 42	0.576 36	0.842 26	0.477 15	0.504 57	0.524 33	0.567 25	0.585 33	0.451 38	0.557 48	1.000 1	0.751 19	0.797 28	0.563 49	1.000 1	0.467 45
Lei Han, Tian Zheng, Lan Xu, Lu Fang: OccuSeg: Occupancy-aware 3D Instance Segmentation. CVPR2020
Mask-Group	0.664 34	1.000 1	0.822 26	0.764 32	0.616 30	0.815 32	0.139 41	0.694 28	0.597 28	0.459 40	0.566 36	0.599 27	0.600 40	0.516 64	0.715 22	0.819 25	0.635 37	1.000 1	0.603 17
Min Zhong, Xinghao Chen, Xiaokang Chen, Gang Zeng, Yunhe Wang: MaskGroup: Hierarchical Point Grouping and Masking for 3D Instance Segmentation. ICME 2022
INS-Conv-instance	0.657 35	1.000 1	0.760 49	0.667 44	0.581 34	0.863 21	0.323 24	0.655 37	0.477 35	0.473 38	0.549 38	0.432 42	0.650 24	1.000 1	0.655 28	0.738 42	0.585 45	0.944 54	0.472 44

CSC-Pretrained	0.648 36	1.000 1	0.810 29	0.768 30	0.523 45	0.813 33	0.143 40	0.819 5	0.389 45	0.422 49	0.511 42	0.443 40	0.650 24	1.000 1	0.624 34	0.732 43	0.634 38	1.000 1	0.375 53

PE	0.645 37	1.000 1	0.773 45	0.798 25	0.538 40	0.786 42	0.088 49	0.799 10	0.350 49	0.435 47	0.547 39	0.545 30	0.646 35	0.933 43	0.562 40	0.761 36	0.556 54	0.997 38	0.501 39
Biao Zhang, Peter Wonka: Point Cloud Instance Segmentation using Probabilistic Embeddings. CVPR 2021
RPGN	0.643 38	1.000 1	0.758 50	0.582 55	0.539 39	0.826 29	0.046 54	0.765 12	0.372 47	0.436 46	0.588 30	0.539 32	0.650 24	1.000 1	0.577 38	0.750 39	0.653 35	0.997 38	0.495 40
Shichao Dong, Guosheng Lin, Tzu-Yi Hung: Learning Regional Purity for Instance Segmentation on 3D Point Clouds. ECCV 2022
Dyco3D	0.641 39	1.000 1	0.841 23	0.893 7	0.531 42	0.802 37	0.115 46	0.588 48	0.448 39	0.438 44	0.537 41	0.430 44	0.550 49	0.857 46	0.534 44	0.764 35	0.657 32	0.987 50	0.568 26
Tong He; Chunhua Shen; Anton van den Hengel: DyCo3D: Robust Instance Segmentation of 3D Point Clouds through Dynamic Convolution. CVPR2021
GICN	0.638 40	1.000 1	0.895 17	0.800 24	0.480 51	0.676 53	0.144 39	0.737 18	0.354 48	0.447 41	0.400 55	0.365 50	0.700 15	1.000 1	0.569 39	0.836 16	0.599 41	1.000 1	0.473 43

PointGroup	0.636 41	1.000 1	0.765 46	0.624 46	0.505 49	0.797 38	0.116 45	0.696 27	0.384 46	0.441 42	0.559 37	0.476 36	0.596 43	1.000 1	0.666 25	0.756 37	0.556 53	0.997 38	0.513 35
Li Jiang, Hengshuang Zhao, Shaoshuai Shi, Shu Liu, Chi-Wing Fu, Jiaya Jia: PointGroup: Dual-Set Point Grouping for 3D Instance Segmentation. CVPR 2020 [oral]
DD-UNet+Group	0.635 42	0.667 57	0.797 39	0.714 39	0.562 37	0.774 43	0.146 38	0.810 8	0.429 41	0.476 37	0.546 40	0.399 46	0.633 36	1.000 1	0.632 33	0.722 45	0.609 40	1.000 1	0.514 34
H. Liu, R. Liu, K. Yang, J. Zhang, K. Peng, R. Stiefelhagen: HIDA: Towards Holistic Indoor Understanding for the Visually Impaired via Semantic Instance Segmentation with a Wearable Solid-State LiDAR Sensor. ICCVW 2021
Mask3D_evaluation	0.631 43	1.000 1	0.829 25	0.606 48	0.646 23	0.836 27	0.068 50	0.511 55	0.462 38	0.507 33	0.619 26	0.389 48	0.610 39	1.000 1	0.432 52	0.828 21	0.673 31	0.788 66	0.552 30

DENet	0.629 44	1.000 1	0.797 38	0.608 47	0.589 33	0.627 57	0.219 33	0.882 1	0.310 51	0.402 54	0.383 57	0.396 47	0.650 24	1.000 1	0.663 27	0.543 65	0.691 29	1.000 1	0.568 27

3D-MPA	0.611 45	1.000 1	0.833 24	0.765 31	0.526 44	0.756 46	0.136 43	0.588 48	0.470 36	0.438 45	0.432 51	0.358 52	0.650 24	0.857 46	0.429 53	0.765 34	0.557 52	1.000 1	0.430 48
Francis Engelmann, Martin Bokeloh, Alireza Fathi, Bastian Leibe, Matthias Nießner: 3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation. CVPR 2020
OSIS	0.605 46	1.000 1	0.801 36	0.599 50	0.535 41	0.728 50	0.286 25	0.436 61	0.679 19	0.491 34	0.433 49	0.256 54	0.404 61	0.857 46	0.620 35	0.724 44	0.510 59	1.000 1	0.539 32

AOIA	0.601 47	1.000 1	0.761 48	0.687 41	0.485 50	0.828 28	0.008 61	0.663 35	0.405 44	0.405 53	0.425 52	0.490 35	0.596 43	0.714 57	0.553 43	0.779 29	0.597 42	0.992 48	0.424 50

PCJC	0.578 48	1.000 1	0.810 30	0.583 54	0.449 54	0.813 34	0.042 55	0.603 46	0.341 50	0.490 35	0.465 46	0.410 45	0.650 24	0.835 54	0.264 63	0.694 52	0.561 50	0.889 59	0.504 38

SSEN	0.575 49	1.000 1	0.761 47	0.473 57	0.477 52	0.795 39	0.066 51	0.529 53	0.658 23	0.460 39	0.461 47	0.380 49	0.331 63	0.859 45	0.401 56	0.692 54	0.653 34	1.000 1	0.348 55
Dongsu Zhang, Junha Chun, Sang Kyun Cha, Young Min Kim: Spatial Semantic Embedding Network: Fast 3D Instance Segmentation with Deep Metric Learning. Arxiv
RWSeg	0.567 50	0.528 67	0.708 58	0.626 45	0.580 35	0.745 49	0.063 52	0.627 42	0.240 55	0.400 55	0.497 43	0.464 37	0.515 50	1.000 1	0.475 48	0.745 40	0.571 47	1.000 1	0.429 49

NeuralBF	0.555 51	0.667 57	0.896 16	0.843 17	0.517 46	0.751 47	0.029 56	0.519 54	0.414 43	0.439 43	0.465 45	0.000 73	0.484 52	0.857 46	0.287 61	0.693 53	0.651 36	1.000 1	0.485 42
Weiwei Sun, Daniel Rebain, Renjie Liao, Vladimir Tankovich, Soroosh Yazdani, Kwang Moo Yi, Andrea Tagliasacchi: NeuralBF: Neural Bilateral Filtering for Top-down Instance Segmentation on Point Clouds. WACV 2023
MTML	0.549 52	1.000 1	0.807 33	0.588 53	0.327 59	0.647 55	0.004 63	0.815 7	0.180 58	0.418 50	0.364 59	0.182 57	0.445 55	1.000 1	0.442 51	0.688 55	0.571 48	1.000 1	0.396 51
Jean Lahoud, Bernard Ghanem, Marc Pollefeys, Martin R. Oswald: 3D Instance Segmentation via Multi-task Metric Learning. ICCV 2019 [oral]
ClickSeg_Instance	0.539 53	1.000 1	0.621 62	0.300 60	0.530 43	0.698 51	0.127 44	0.533 52	0.222 56	0.430 48	0.400 54	0.365 50	0.574 47	0.938 42	0.472 49	0.659 57	0.543 55	0.944 54	0.347 56

One_Thing_One_Click	0.529 54	0.667 57	0.718 54	0.777 28	0.399 55	0.683 52	0.000 66	0.669 33	0.138 61	0.391 56	0.374 58	0.539 31	0.360 62	0.641 61	0.556 42	0.774 32	0.593 43	0.997 38	0.251 61
Zhengzhe Liu, Xiaojuan Qi, Chi-Wing Fu: One Thing One Click: A Self-Training Approach for Weakly Supervised 3D Semantic Segmentation. CVPR 2021
Sparse R-CNN	0.515 55	1.000 1	0.538 67	0.282 61	0.468 53	0.790 40	0.173 35	0.345 63	0.429 40	0.413 52	0.484 44	0.176 58	0.595 45	0.591 62	0.522 45	0.668 56	0.476 60	0.986 52	0.327 57

Occipital-SCS	0.512 56	1.000 1	0.716 55	0.509 56	0.506 48	0.611 58	0.092 48	0.602 47	0.177 59	0.346 59	0.383 56	0.165 59	0.442 56	0.850 53	0.386 58	0.618 61	0.543 56	0.889 59	0.389 52

3D-BoNet	0.488 57	1.000 1	0.672 61	0.590 52	0.301 61	0.484 68	0.098 47	0.620 44	0.306 52	0.341 60	0.259 63	0.125 61	0.434 58	0.796 56	0.402 55	0.499 67	0.513 58	0.909 58	0.439 47
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-inst	0.478 58	0.667 57	0.712 57	0.595 51	0.259 64	0.550 64	0.000 66	0.613 45	0.175 60	0.250 65	0.434 48	0.437 41	0.411 60	0.857 46	0.485 47	0.591 64	0.267 70	0.944 54	0.359 54
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
SPG_WSIS	0.470 59	0.667 57	0.685 60	0.677 43	0.372 57	0.562 62	0.000 66	0.482 58	0.244 54	0.316 62	0.298 60	0.052 68	0.442 57	0.857 46	0.267 62	0.702 49	0.559 51	1.000 1	0.287 59

SALoss-ResNet	0.459 60	1.000 1	0.737 53	0.159 71	0.259 63	0.587 60	0.138 42	0.475 59	0.217 57	0.416 51	0.408 53	0.128 60	0.315 64	0.714 57	0.411 54	0.536 66	0.590 44	0.873 63	0.304 58
Zhidong Liang, Ming Yang, Hao Li, Chunxiang Wang: 3D Instance Embedding Learning With a Structure-Aware Loss Function for Point Cloud Segmentation. IEEE Robotics and Automation Letters (IROS2020)
MASC	0.447 61	0.528 67	0.555 65	0.381 58	0.382 56	0.633 56	0.002 64	0.509 56	0.260 53	0.361 58	0.432 50	0.327 53	0.451 54	0.571 63	0.367 59	0.639 59	0.386 61	0.980 53	0.276 60
Chen Liu, Yasutaka Furukawa: MASC: Multi-scale Affinity with Sparse Convolution for 3D Instance Segmentation.
SegGroup_ins	0.445 62	0.667 57	0.773 44	0.185 68	0.317 60	0.656 54	0.000 66	0.407 62	0.134 62	0.381 57	0.267 62	0.217 56	0.476 53	0.714 57	0.452 50	0.629 60	0.514 57	1.000 1	0.222 64
An Tao, Yueqi Duan, Yi Wei, Jiwen Lu, Jie Zhou: SegGroup: Seg-Level Supervision for 3D Instance and Semantic Segmentation. TIP 2022
3D-SIS	0.382 63	1.000 1	0.432 70	0.245 63	0.190 65	0.577 61	0.013 60	0.263 65	0.033 68	0.320 61	0.240 64	0.075 64	0.422 59	0.857 46	0.117 68	0.699 50	0.271 69	0.883 61	0.235 63
Ji Hou, Angela Dai, Matthias Niessner: 3D-SIS: 3D Semantic Instance Segmentation of RGB-D Scans. CVPR 2019
Hier3D	0.323 64	0.667 57	0.542 66	0.264 62	0.157 68	0.550 63	0.000 66	0.205 68	0.009 70	0.270 64	0.218 65	0.075 64	0.500 51	0.688 60	0.007 74	0.698 51	0.301 66	0.459 71	0.200 65
Tan: HCFS3D: Hierarchical Coupled Feature Selection Network for 3D Semantic and Instance Segmentation.
UNet-backbone	0.319 65	0.667 57	0.715 56	0.233 64	0.189 66	0.479 69	0.008 61	0.218 66	0.067 67	0.201 67	0.173 66	0.107 62	0.123 69	0.438 65	0.150 65	0.615 62	0.355 62	0.916 57	0.093 73

R-PointNet	0.306 66	0.500 69	0.405 71	0.311 59	0.348 58	0.589 59	0.054 53	0.068 71	0.126 63	0.283 63	0.290 61	0.028 69	0.219 67	0.214 68	0.331 60	0.396 71	0.275 67	0.821 65	0.245 62

Region-18class	0.284 67	0.250 73	0.751 52	0.228 66	0.270 62	0.521 65	0.000 66	0.468 60	0.008 72	0.205 66	0.127 67	0.000 73	0.068 71	0.070 72	0.262 64	0.652 58	0.323 64	0.740 67	0.173 66

SemRegionNet-20cls	0.250 68	0.333 70	0.613 63	0.229 65	0.163 67	0.493 66	0.000 66	0.304 64	0.107 64	0.147 70	0.100 69	0.052 67	0.231 65	0.119 70	0.039 70	0.445 69	0.325 63	0.654 68	0.141 69

3D-BEVIS	0.248 69	0.667 57	0.566 64	0.076 72	0.035 74	0.394 72	0.027 58	0.035 73	0.098 65	0.099 72	0.030 73	0.025 70	0.098 70	0.375 67	0.126 67	0.604 63	0.181 72	0.854 64	0.171 67
Cathrin Elich, Francis Engelmann, Jonas Schult, Theodora Kontogianni, Bastian Leibe: 3D-BEVIS: Birds-Eye-View Instance Segmentation.
tmp	0.248 69	0.667 57	0.437 69	0.188 67	0.153 69	0.491 67	0.000 66	0.208 67	0.094 66	0.153 69	0.099 70	0.057 66	0.217 68	0.119 70	0.039 70	0.466 68	0.302 65	0.640 69	0.140 70

Sem_Recon_ins	0.227 71	0.764 56	0.486 68	0.069 73	0.098 71	0.426 71	0.017 59	0.067 72	0.015 69	0.172 68	0.100 68	0.096 63	0.054 73	0.183 69	0.135 66	0.366 72	0.260 71	0.614 70	0.168 68

ASIS	0.199 72	0.333 70	0.253 73	0.167 70	0.140 70	0.438 70	0.000 66	0.177 69	0.008 71	0.121 71	0.069 71	0.004 72	0.231 66	0.429 66	0.036 72	0.445 70	0.273 68	0.333 73	0.119 72

Sgpn_scannet	0.143 73	0.208 74	0.390 72	0.169 69	0.065 72	0.275 73	0.029 57	0.069 70	0.000 73	0.087 73	0.043 72	0.014 71	0.027 74	0.000 73	0.112 69	0.351 73	0.168 73	0.438 72	0.138 71

MaskRCNN 2d->3d Proj	0.058 74	0.333 70	0.002 74	0.000 74	0.053 73	0.002 74	0.002 65	0.021 74	0.000 73	0.045 74	0.024 74	0.238 55	0.065 72	0.000 73	0.014 73	0.107 74	0.020 74	0.110 74	0.006 74

3D Semantic Instance Benchmark