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	bathtub	bed	bookshelf	cabinet	chair	counter	curtain	desk	door	otherfurniture	picture	refrigerator	shower curtain	sink	sofa	table	toilet	window

Volt-SPFormerScanNet	0.640 1	0.920 12	0.665 12	0.634 1	0.582 1	0.794 5	0.242 19	0.559 21	0.496 1	0.535 2	0.646 1	0.737 2	0.709 6	0.731 29	0.509 4	0.796 2	0.566 16	0.944 10	0.457 2
Kadir Yilmaz, Adrian Kruse, Tristan Höfer, Daan de Geus, Bastian Leibe: Volume Transformer: Revisiting Vanilla Transformers for 3D Scene Understanding.
PointComp	0.629 2	0.787 26	0.679 10	0.574 6	0.502 4	0.824 1	0.378 1	0.480 40	0.483 4	0.480 16	0.601 2	0.744 1	0.682 9	0.809 8	0.460 21	0.819 1	0.643 2	0.935 14	0.449 4

PointRel	0.622 3	0.926 7	0.710 3	0.541 12	0.502 3	0.772 10	0.314 5	0.598 11	0.425 11	0.504 12	0.565 4	0.650 9	0.716 2	0.809 7	0.476 13	0.747 7	0.618 3	0.963 3	0.364 22
: Relation3D: Enhancing Relation Modeling for Point Cloud Instance Segmentation. CVPR 2025
Competitor-MAFT	0.618 4	0.866 16	0.724 1	0.628 2	0.484 6	0.803 3	0.300 9	0.509 33	0.496 2	0.539 1	0.547 8	0.703 3	0.668 10	0.708 35	0.463 18	0.708 20	0.595 4	0.959 5	0.418 9

SIM3D	0.617 5	0.952 3	0.629 20	0.539 13	0.426 18	0.768 14	0.302 8	0.681 2	0.425 12	0.473 18	0.511 18	0.701 4	0.717 1	0.821 6	0.467 16	0.774 3	0.559 17	0.914 21	0.448 5

Spherical Mask(CtoF)	0.616 6	0.946 4	0.654 15	0.555 8	0.434 15	0.769 13	0.271 14	0.604 8	0.447 6	0.505 10	0.549 5	0.698 5	0.716 2	0.775 17	0.480 10	0.747 8	0.575 12	0.925 16	0.436 7

EV3D	0.615 7	0.946 4	0.652 16	0.555 8	0.433 16	0.773 9	0.271 15	0.604 8	0.447 6	0.506 9	0.544 9	0.698 5	0.716 2	0.775 17	0.480 10	0.747 8	0.572 14	0.925 16	0.435 8

DCD	0.614 8	0.892 13	0.633 19	0.434 30	0.495 5	0.810 2	0.292 10	0.501 34	0.408 13	0.525 6	0.582 3	0.688 7	0.625 12	0.801 9	0.608 1	0.672 23	0.649 1	0.965 2	0.476 1

ExtMask3D	0.598 9	0.852 17	0.692 8	0.433 33	0.461 9	0.791 6	0.264 16	0.488 37	0.493 3	0.508 8	0.528 17	0.594 15	0.706 7	0.791 11	0.483 8	0.734 12	0.595 5	0.911 23	0.437 6

MAFT	0.596 10	0.889 14	0.721 2	0.448 25	0.460 10	0.768 15	0.251 18	0.558 22	0.408 14	0.504 11	0.539 11	0.616 12	0.618 14	0.858 3	0.482 9	0.684 22	0.551 20	0.931 15	0.450 3

MG-Former	0.587 11	0.852 17	0.639 18	0.454 24	0.393 24	0.758 18	0.338 3	0.572 16	0.480 5	0.527 4	0.491 25	0.671 8	0.527 24	0.867 1	0.485 7	0.601 34	0.590 8	0.938 13	0.390 14

InsSSM	0.586 12	1.000 1	0.593 24	0.440 28	0.480 7	0.771 11	0.345 2	0.437 43	0.444 9	0.495 15	0.548 7	0.579 19	0.621 13	0.720 31	0.409 25	0.712 15	0.593 6	0.960 4	0.395 12
Lei Yao, Yi Wang, Moyun Liu, Lap-Pui Chau: SGIFormer: Semantic-guided and Geometric-enhanced Interleaving Transformer for 3D Instance Segmentation. TCSVT, 2024
Queryformer	0.583 13	0.926 7	0.702 5	0.393 39	0.504 2	0.733 24	0.276 13	0.527 28	0.373 20	0.479 17	0.534 13	0.533 26	0.697 8	0.720 32	0.436 23	0.745 10	0.592 7	0.958 6	0.363 23

KmaxOneFormerNet	0.581 14	0.745 31	0.692 9	0.551 10	0.458 11	0.798 4	0.264 17	0.531 27	0.369 22	0.513 7	0.531 16	0.632 10	0.494 27	0.798 10	0.567 3	0.648 27	0.558 19	0.950 8	0.362 25

Competitor-SPFormer	0.580 15	0.721 38	0.705 4	0.593 5	0.444 14	0.786 8	0.286 11	0.564 19	0.376 19	0.498 14	0.534 14	0.546 24	0.390 48	0.785 13	0.577 2	0.708 19	0.579 10	0.954 7	0.388 15

VDG-Uni3DSeg	0.576 16	0.833 21	0.699 6	0.483 18	0.412 22	0.767 16	0.313 6	0.461 42	0.446 8	0.526 5	0.498 23	0.584 16	0.551 20	0.743 26	0.464 17	0.766 4	0.538 24	0.919 19	0.363 24

PBNet	0.573 17	0.926 7	0.575 30	0.619 3	0.472 8	0.736 22	0.239 21	0.487 38	0.383 18	0.459 22	0.506 21	0.533 25	0.585 16	0.767 19	0.404 27	0.717 14	0.559 18	0.969 1	0.381 18
Weiguang Zhao, Yuyao Yan, Chaolong Yang, Jianan Ye, Xi Yang, Kaizhu Huang: Divide and Conquer: 3D Instance Segmentation With Point-Wise Binarization. ICCV 2023
TST3D	0.569 18	0.778 28	0.675 11	0.598 4	0.451 13	0.727 25	0.280 12	0.476 41	0.395 15	0.472 19	0.457 31	0.583 17	0.580 18	0.777 14	0.462 20	0.735 11	0.547 22	0.919 20	0.333 31
Duc Tran Dang Trung, Byeongkeun Kang, Yeejin Lee: MSTA3D: Multi-scale Twin-attention for 3D Instance Segmentation. ACM Multimedia 2024
Mask3D	0.566 19	0.926 7	0.597 23	0.408 36	0.420 20	0.737 21	0.239 20	0.598 11	0.386 17	0.458 23	0.549 5	0.568 22	0.716 2	0.601 48	0.480 10	0.646 28	0.575 12	0.922 18	0.364 21
Jonas Schult, Francis Engelmann, Alexander Hermans, Or Litany, Siyu Tang, Bastian Leibe: Mask3D for 3D Semantic Instance Segmentation. ICRA 2023
OneFormer3D	0.566 19	0.781 27	0.697 7	0.562 7	0.431 17	0.770 12	0.331 4	0.400 49	0.373 21	0.529 3	0.504 22	0.568 21	0.475 31	0.732 28	0.470 14	0.762 5	0.550 21	0.871 38	0.379 19
Maxim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: OneFormer3D: One Transformer for Unified Point Cloud Segmentation.
UniPerception	0.560 21	0.815 22	0.659 13	0.388 41	0.453 12	0.786 7	0.212 22	0.526 29	0.441 10	0.471 20	0.539 10	0.607 13	0.442 36	0.671 42	0.406 26	0.731 13	0.577 11	0.944 11	0.411 10

ISBNet	0.559 22	0.939 6	0.655 14	0.383 43	0.426 19	0.763 17	0.180 24	0.534 26	0.386 16	0.499 13	0.509 20	0.621 11	0.427 42	0.704 37	0.467 15	0.649 26	0.571 15	0.948 9	0.401 11
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
GraphCut	0.552 23	1.000 1	0.611 22	0.438 29	0.392 25	0.714 26	0.139 28	0.598 13	0.327 26	0.389 26	0.510 19	0.598 14	0.427 43	0.754 22	0.463 19	0.761 6	0.588 9	0.903 26	0.329 33

SPFormer	0.549 24	0.745 31	0.640 17	0.484 17	0.395 23	0.739 20	0.311 7	0.566 18	0.335 24	0.468 21	0.492 24	0.555 23	0.478 30	0.747 24	0.436 22	0.712 16	0.540 23	0.893 30	0.343 30
Sun Jiahao, Qing Chunmei, Tan Junpeng, Xu Xiangmin: Superpoint Transformer for 3D Scene Instance Segmentation. AAAI 2023 [Oral]
DKNet	0.532 25	0.815 22	0.624 21	0.517 14	0.377 27	0.749 19	0.107 30	0.509 32	0.304 28	0.437 24	0.475 26	0.581 18	0.539 22	0.775 16	0.339 33	0.640 30	0.506 27	0.901 27	0.385 17
Yizheng Wu, Min Shi, Shuaiyuan Du, Hao Lu, Zhiguo Cao, Weicai Zhong: 3D Instances as 1D Kernels. ECCV 2022
IPCA-Inst	0.520 26	0.889 14	0.551 34	0.548 11	0.418 21	0.665 36	0.064 39	0.585 14	0.260 36	0.277 41	0.471 28	0.500 27	0.644 11	0.785 12	0.369 29	0.591 38	0.511 25	0.878 35	0.362 26

SoftGroup++	0.513 27	0.704 40	0.578 29	0.398 38	0.363 33	0.704 27	0.061 40	0.647 5	0.297 33	0.378 29	0.537 12	0.343 31	0.614 15	0.828 5	0.295 38	0.710 18	0.505 29	0.875 37	0.394 13

SSTNet	0.506 28	0.738 35	0.549 35	0.497 16	0.316 39	0.693 30	0.178 25	0.377 53	0.198 42	0.330 32	0.463 30	0.576 20	0.515 25	0.857 4	0.494 5	0.637 31	0.457 33	0.943 12	0.290 42
Zhihao Liang, Zhihao Li, Songcen Xu, Mingkui Tan, Kui Jia: Instance Segmentation in 3D Scenes using Semantic Superpoint Tree Networks. ICCV2021
DANCENET	0.504 29	0.926 7	0.579 26	0.472 20	0.367 30	0.626 46	0.165 26	0.432 44	0.221 38	0.408 25	0.449 33	0.411 29	0.564 19	0.746 25	0.421 24	0.707 21	0.438 36	0.846 46	0.288 43

SoftGroup	0.504 29	0.667 47	0.579 27	0.372 45	0.381 26	0.694 29	0.072 36	0.677 3	0.303 29	0.387 27	0.531 15	0.319 35	0.582 17	0.754 21	0.318 34	0.643 29	0.492 30	0.907 25	0.388 16
Thang Vu, Kookhoi Kim, Tung M. Luu, Xuan Thanh Nguyen, Chang D. Yoo: SoftGroup for 3D Instance Segmentaiton on Point Clouds. CVPR 2022 [Oral]
TD3D	0.489 31	0.852 17	0.511 44	0.434 31	0.322 38	0.735 23	0.101 33	0.512 31	0.355 23	0.349 31	0.468 29	0.283 39	0.514 26	0.676 41	0.268 43	0.671 24	0.510 26	0.908 24	0.329 34
Maksim Kolodiazhnyi, Anna Vorontsova, Anton Konushin, Danila Rukhovich: Top-Down Beats Bottom-Up in 3D Instance Segmentation. WACV 2024
OccuSeg+instance	0.486 32	0.802 25	0.536 37	0.428 34	0.369 29	0.702 28	0.205 23	0.331 58	0.301 30	0.379 28	0.474 27	0.327 32	0.437 37	0.862 2	0.485 6	0.601 35	0.394 44	0.846 48	0.273 46
Lei Han, Tian Zheng, Lan Xu, Lu Fang: OccuSeg: Occupancy-aware 3D Instance Segmentation. CVPR2020
TopoSeg	0.479 33	0.704 40	0.564 31	0.467 22	0.366 31	0.633 44	0.068 37	0.554 23	0.262 35	0.328 33	0.447 34	0.323 33	0.534 23	0.722 30	0.288 40	0.614 32	0.482 31	0.912 22	0.358 28

DualGroup	0.469 34	0.815 22	0.552 33	0.398 37	0.374 28	0.683 32	0.130 29	0.539 25	0.310 27	0.327 34	0.407 37	0.276 40	0.447 35	0.535 52	0.342 32	0.659 25	0.455 34	0.900 29	0.301 38

SSEC	0.465 35	0.667 47	0.578 28	0.502 15	0.362 34	0.641 43	0.035 49	0.605 7	0.291 34	0.323 35	0.451 32	0.296 37	0.417 46	0.677 40	0.245 47	0.501 56	0.506 28	0.900 28	0.366 20

ODIN - Ins	0.463 36	0.738 35	0.589 25	0.344 49	0.358 35	0.560 55	0.139 27	0.393 52	0.331 25	0.373 30	0.392 40	0.496 28	0.493 28	0.709 34	0.377 28	0.599 36	0.359 50	0.752 58	0.332 32
Ayush Jain, Pushkal Katara, Nikolaos Gkanatsios, Adam W. Harley, Gabriel Sarch, Kriti Aggarwal, Vishrav Chaudhary, Katerina Fragkiadaki: ODIN: A Single Model for 2D and 3D Segmentation. CVPR 2024
HAIS	0.457 37	0.704 40	0.561 32	0.457 23	0.364 32	0.673 33	0.046 48	0.547 24	0.194 43	0.308 36	0.426 35	0.288 38	0.454 34	0.711 33	0.262 44	0.563 46	0.434 38	0.889 32	0.344 29
Shaoyu Chen, Jiemin Fang, Qian Zhang, Wenyu Liu, Xinggang Wang: Hierarchical Aggregation for 3D Instance Segmentation. ICCV 2021
DD-UNet+Group	0.436 38	0.630 55	0.508 47	0.480 19	0.310 41	0.624 48	0.065 38	0.638 6	0.174 44	0.256 45	0.384 42	0.194 52	0.428 40	0.759 20	0.289 39	0.574 43	0.400 42	0.849 45	0.291 41
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
INS-Conv-instance	0.435 39	0.716 39	0.495 49	0.355 47	0.331 36	0.689 31	0.102 32	0.394 51	0.208 41	0.280 39	0.395 39	0.250 43	0.544 21	0.741 27	0.309 36	0.536 52	0.391 45	0.842 51	0.258 50

Mask-Group	0.434 40	0.778 28	0.516 42	0.471 21	0.330 37	0.658 37	0.029 51	0.526 30	0.249 37	0.256 44	0.400 38	0.309 36	0.384 51	0.296 68	0.368 30	0.575 42	0.425 39	0.877 36	0.362 27
Min Zhong, Xinghao Chen, Xiaokang Chen, Gang Zeng, Yunhe Wang: MaskGroup: Hierarchical Point Grouping and Masking for 3D Instance Segmentation. ICME 2022
Box2Mask	0.433 41	0.741 33	0.463 54	0.433 32	0.283 44	0.625 47	0.103 31	0.298 63	0.125 53	0.260 43	0.424 36	0.322 34	0.472 32	0.701 38	0.363 31	0.711 17	0.309 62	0.882 33	0.272 48
Julian Chibane, Francis Engelmann, Tuan Anh Tran, Gerard Pons-Moll: Box2Mask: Weakly Supervised 3D Semantic Instance Segmentation Using Bounding Boxes. ECCV 2022
RPGN	0.428 42	0.630 55	0.508 46	0.367 46	0.249 51	0.658 38	0.016 59	0.673 4	0.131 51	0.234 48	0.383 43	0.270 41	0.434 38	0.748 23	0.274 42	0.609 33	0.406 41	0.842 50	0.267 49
Shichao Dong, Guosheng Lin, Tzu-Yi Hung: Learning Regional Purity for Instance Segmentation on 3D Point Clouds. ECCV 2022
DENet	0.413 43	0.741 33	0.520 39	0.237 59	0.284 43	0.523 58	0.097 34	0.691 1	0.138 48	0.209 58	0.229 60	0.238 46	0.390 49	0.707 36	0.310 35	0.448 63	0.470 32	0.892 31	0.310 36

PointGroup	0.407 44	0.639 54	0.496 48	0.415 35	0.243 53	0.645 42	0.021 56	0.570 17	0.114 54	0.211 56	0.359 45	0.217 50	0.428 41	0.660 43	0.256 45	0.562 47	0.341 54	0.860 41	0.291 40
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]
CSC-Pretrained	0.405 45	0.738 35	0.465 53	0.331 52	0.205 57	0.655 39	0.051 44	0.601 10	0.092 58	0.211 57	0.329 48	0.198 51	0.459 33	0.775 15	0.195 54	0.524 54	0.400 43	0.878 34	0.184 59

PE	0.396 46	0.667 47	0.467 52	0.446 27	0.243 52	0.624 49	0.022 55	0.577 15	0.106 55	0.219 51	0.340 46	0.239 45	0.487 29	0.475 59	0.225 49	0.541 51	0.350 52	0.818 53	0.273 47
Biao Zhang, Peter Wonka: Point Cloud Instance Segmentation using Probabilistic Embeddings. CVPR 2021
Dyco3D	0.395 47	0.642 53	0.518 41	0.447 26	0.259 50	0.666 35	0.050 45	0.251 68	0.166 45	0.231 49	0.362 44	0.232 47	0.331 54	0.535 51	0.229 48	0.587 39	0.438 37	0.850 43	0.317 35
Tong He; Chunhua Shen; Anton van den Hengel: DyCo3D: Robust Instance Segmentation of 3D Point Clouds through Dynamic Convolution. CVPR2021
OSIS	0.392 48	0.778 28	0.530 38	0.220 61	0.278 45	0.567 54	0.083 35	0.330 59	0.299 31	0.270 42	0.310 51	0.143 58	0.260 58	0.624 46	0.277 41	0.568 45	0.361 49	0.865 40	0.301 37

AOIA	0.387 49	0.704 40	0.515 43	0.385 42	0.225 56	0.669 34	0.005 66	0.482 39	0.126 52	0.181 61	0.269 57	0.221 49	0.426 44	0.478 58	0.218 50	0.592 37	0.371 47	0.851 42	0.242 52

SSEN	0.384 50	0.852 17	0.494 50	0.192 62	0.226 55	0.648 41	0.022 54	0.398 50	0.299 32	0.277 40	0.317 50	0.231 48	0.194 65	0.514 55	0.196 52	0.586 40	0.444 35	0.843 49	0.184 58
Dongsu Zhang, Junha Chun, Sang Kyun Cha, Young Min Kim: Spatial Semantic Embedding Network: Fast 3D Instance Segmentation with Deep Metric Learning. Arxiv
Mask3D_evaluation	0.382 51	0.593 57	0.520 40	0.390 40	0.314 40	0.600 50	0.018 58	0.287 66	0.151 47	0.281 38	0.387 41	0.169 56	0.429 39	0.654 44	0.172 58	0.578 41	0.384 46	0.670 65	0.278 45

PCJC	0.375 52	0.704 40	0.542 36	0.284 56	0.197 59	0.649 40	0.006 63	0.426 45	0.138 49	0.242 46	0.304 52	0.183 55	0.388 50	0.629 45	0.141 65	0.546 50	0.344 53	0.738 60	0.283 44

ClickSeg_Instance	0.366 53	0.654 51	0.375 58	0.184 63	0.302 42	0.592 52	0.050 46	0.300 62	0.093 57	0.283 37	0.277 54	0.249 44	0.426 45	0.615 47	0.299 37	0.504 55	0.367 48	0.832 52	0.191 57

SphereSeg	0.357 54	0.651 52	0.411 56	0.345 48	0.264 49	0.630 45	0.059 41	0.289 65	0.212 39	0.240 47	0.336 47	0.158 57	0.305 55	0.557 49	0.159 61	0.455 62	0.341 55	0.726 62	0.294 39

3D-MPA	0.355 55	0.457 67	0.484 51	0.299 54	0.277 46	0.591 53	0.047 47	0.332 56	0.212 40	0.217 52	0.278 53	0.193 53	0.413 47	0.410 62	0.195 53	0.574 44	0.352 51	0.849 44	0.213 55
Francis Engelmann, Martin Bokeloh, Alireza Fathi, Bastian Leibe, Matthias Nießner: 3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation. CVPR 2020
NeuralBF	0.353 56	0.593 57	0.511 45	0.375 44	0.264 48	0.597 51	0.008 61	0.332 57	0.160 46	0.229 50	0.274 56	0.000 79	0.206 62	0.678 39	0.155 62	0.485 58	0.422 40	0.816 54	0.254 51
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
RWSeg	0.348 57	0.475 64	0.456 55	0.320 53	0.275 47	0.476 60	0.020 57	0.491 36	0.056 65	0.212 55	0.320 49	0.261 42	0.302 56	0.520 53	0.182 56	0.557 48	0.285 64	0.867 39	0.197 56

GICN	0.341 58	0.580 59	0.371 59	0.344 50	0.198 58	0.469 61	0.052 43	0.564 20	0.093 56	0.212 54	0.212 62	0.127 60	0.347 53	0.537 50	0.206 51	0.525 53	0.329 57	0.729 61	0.241 53

One_Thing_One_Click	0.326 59	0.472 65	0.361 60	0.232 60	0.183 60	0.555 56	0.000 72	0.498 35	0.038 67	0.195 59	0.226 61	0.362 30	0.168 66	0.469 60	0.251 46	0.553 49	0.335 56	0.846 47	0.117 67
Zhengzhe Liu, Xiaojuan Qi, Chi-Wing Fu: One Thing One Click: A Self-Training Approach for Weakly Supervised 3D Semantic Segmentation. CVPR 2021
Occipital-SCS	0.320 60	0.679 46	0.352 61	0.334 51	0.229 54	0.436 62	0.025 52	0.412 48	0.058 63	0.161 66	0.240 59	0.085 62	0.262 57	0.496 57	0.187 55	0.467 60	0.328 58	0.775 55	0.231 54

Sparse R-CNN	0.292 61	0.704 40	0.213 71	0.153 65	0.154 62	0.551 57	0.053 42	0.212 69	0.132 50	0.174 63	0.274 55	0.070 64	0.363 52	0.441 61	0.176 57	0.424 65	0.234 66	0.758 57	0.161 63

MTML	0.282 62	0.577 60	0.380 57	0.182 64	0.107 68	0.430 63	0.001 69	0.422 46	0.057 64	0.179 62	0.162 65	0.070 65	0.229 60	0.511 56	0.161 59	0.491 57	0.313 59	0.650 68	0.162 61
Jean Lahoud, Bernard Ghanem, Marc Pollefeys, Martin R. Oswald: 3D Instance Segmentation via Multi-task Metric Learning. ICCV 2019 [oral]
SALoss-ResNet	0.262 63	0.667 47	0.335 62	0.067 72	0.123 66	0.427 64	0.022 53	0.280 67	0.058 62	0.216 53	0.211 63	0.039 68	0.142 68	0.519 54	0.106 69	0.338 69	0.310 61	0.721 63	0.138 64
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	0.254 64	0.463 66	0.249 70	0.113 66	0.167 61	0.412 66	0.000 71	0.374 54	0.073 59	0.173 64	0.243 58	0.130 59	0.228 61	0.368 64	0.160 60	0.356 67	0.208 67	0.711 64	0.136 65
Chen Liu, Yasutaka Furukawa: MASC: Multi-scale Affinity with Sparse Convolution for 3D Instance Segmentation.
3D-BoNet	0.253 65	0.519 62	0.324 65	0.251 58	0.137 65	0.345 71	0.031 50	0.419 47	0.069 60	0.162 65	0.131 67	0.052 66	0.202 64	0.338 66	0.147 64	0.301 72	0.303 63	0.651 67	0.178 60
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
SPG_WSIS	0.251 66	0.380 69	0.274 68	0.289 55	0.144 63	0.413 65	0.000 72	0.311 60	0.065 61	0.113 68	0.130 68	0.029 71	0.204 63	0.388 63	0.108 68	0.459 61	0.311 60	0.769 56	0.127 66

SegGroup_ins	0.246 67	0.556 61	0.335 63	0.062 74	0.115 67	0.490 59	0.000 72	0.297 64	0.018 71	0.186 60	0.142 66	0.083 63	0.233 59	0.216 70	0.153 63	0.469 59	0.251 65	0.744 59	0.083 70
An Tao, Yueqi Duan, Yi Wei, Jiwen Lu, Jie Zhou: SegGroup: Seg-Level Supervision for 3D Instance and Semantic Segmentation. TIP 2022
PanopticFusion-inst	0.214 68	0.250 74	0.330 64	0.275 57	0.103 69	0.228 77	0.000 72	0.345 55	0.024 69	0.088 70	0.203 64	0.186 54	0.167 67	0.367 65	0.125 66	0.221 75	0.112 77	0.666 66	0.162 62
Gaku Narita, Takashi Seno, Tomoya Ishikawa, Yohsuke Kaji: PanopticFusion: Online Volumetric Semantic Mapping at the Level of Stuff and Things. IROS 2019 (to appear)
UNet-backbone	0.161 69	0.519 62	0.259 69	0.084 68	0.059 71	0.325 73	0.002 67	0.093 74	0.009 73	0.077 72	0.064 71	0.045 67	0.044 75	0.161 72	0.045 71	0.331 70	0.180 69	0.566 69	0.033 79

3D-SIS	0.161 69	0.407 68	0.155 76	0.068 71	0.043 75	0.346 70	0.001 68	0.134 71	0.005 74	0.088 69	0.106 70	0.037 69	0.135 70	0.321 67	0.028 75	0.339 68	0.116 76	0.466 72	0.093 69
Ji Hou, Angela Dai, Matthias Niessner: 3D-SIS: 3D Semantic Instance Segmentation of RGB-D Scans. CVPR 2019
R-PointNet	0.158 71	0.356 70	0.173 74	0.113 67	0.140 64	0.359 67	0.012 60	0.023 77	0.039 66	0.134 67	0.123 69	0.008 75	0.089 71	0.149 73	0.117 67	0.221 74	0.128 74	0.563 70	0.094 68

Region-18class	0.146 72	0.175 78	0.321 66	0.080 69	0.062 70	0.357 68	0.000 72	0.307 61	0.002 76	0.066 73	0.044 73	0.000 79	0.018 77	0.036 78	0.054 70	0.447 64	0.133 72	0.472 71	0.060 74

SemRegionNet-20cls	0.121 73	0.296 72	0.203 72	0.071 70	0.058 72	0.349 69	0.000 72	0.150 70	0.019 70	0.054 75	0.034 76	0.017 74	0.052 73	0.042 77	0.013 78	0.209 76	0.183 68	0.371 73	0.057 75

3D-BEVIS	0.117 74	0.250 74	0.308 67	0.020 78	0.009 80	0.269 76	0.006 64	0.008 78	0.029 68	0.037 78	0.014 79	0.003 77	0.036 76	0.147 74	0.042 73	0.381 66	0.118 75	0.362 74	0.069 73
Cathrin Elich, Francis Engelmann, Jonas Schult, Theodora Kontogianni, Bastian Leibe: 3D-BEVIS: Birds-Eye-View Instance Segmentation.
Hier3D	0.117 74	0.222 76	0.161 75	0.054 76	0.027 77	0.289 74	0.000 72	0.124 72	0.001 78	0.079 71	0.061 72	0.027 72	0.141 69	0.240 69	0.005 79	0.310 71	0.129 73	0.153 79	0.081 71
Tan: HCFS3D: Hierarchical Coupled Feature Selection Network for 3D Semantic and Instance Segmentation.
tmp	0.113 76	0.333 71	0.151 77	0.056 75	0.053 73	0.344 72	0.000 72	0.105 73	0.016 72	0.049 76	0.035 75	0.020 73	0.053 72	0.048 76	0.013 77	0.183 78	0.173 70	0.344 76	0.054 76

Sem_Recon_ins	0.098 77	0.295 73	0.187 73	0.015 79	0.036 76	0.213 78	0.005 65	0.038 76	0.003 75	0.056 74	0.037 74	0.036 70	0.015 78	0.051 75	0.044 72	0.209 77	0.098 78	0.354 75	0.071 72

ASIS	0.085 78	0.037 79	0.080 79	0.066 73	0.047 74	0.282 75	0.000 72	0.052 75	0.002 77	0.047 77	0.026 77	0.001 78	0.046 74	0.194 71	0.031 74	0.264 73	0.140 71	0.167 78	0.047 78

Sgpn_scannet	0.049 79	0.023 80	0.134 78	0.031 77	0.013 79	0.144 79	0.006 62	0.008 79	0.000 79	0.028 79	0.017 78	0.003 76	0.009 80	0.000 79	0.021 76	0.122 79	0.095 79	0.175 77	0.054 77

MaskRCNN 2d->3d Proj	0.022 80	0.185 77	0.000 80	0.000 80	0.015 78	0.000 80	0.000 70	0.006 80	0.000 79	0.010 80	0.006 80	0.107 61	0.012 79	0.000 79	0.002 80	0.027 80	0.004 80	0.022 80	0.001 80

3D Semantic Instance Benchmark