This table lists the benchmark results for the ScanRefer Localization Benchmark scenario.


   Unique Unique Multiple Multiple Overall Overall
Method Infoacc@0.25IoUacc@0.5IoUacc@0.25IoUacc@0.5IoUacc@0.25IoUacc@0.5IoU
sort bysort bysort bysort bysort bysorted by
3DVG-Transformer+0.7733 20.5787 40.4370 10.3102 10.5124 10.3704 1
InstanceReferpermissive0.7782 10.6669 10.3457 50.2688 30.4427 40.3580 2
Zhihao Yuan, Xu Yan, Yinghong Liao, Ruimao Zhang, Zhen Li*, Shuguang Cui: InstanceRefer: Cooperative Holistic Understanding for Visual Grounding on Point Clouds through Instance Multi-level Contextual Referring. arXiv preprint
3DVG-Transformer0.7576 30.5515 50.4224 20.2933 20.4976 20.3512 3
PointGroup_MCAN0.7510 40.6397 20.3271 70.2535 40.4222 60.3401 4
TGNN0.6834 70.5894 30.3312 60.2526 50.4102 70.3281 5
Pin-Hao Huang, Han-Hung Lee, Hwann-Tzong Chen, Tyng-Luh Liu: Text-Guided Graph Neural Network for Referring 3D Instance Segmentation. AAAI 2021
SRGA0.7494 50.5128 60.3631 30.2218 60.4497 30.2871 6
ScanReferpermissive0.6859 60.4353 70.3488 40.2097 70.4244 50.2603 7
Dave Zhenyu Chen, Angel X. Chang, Matthias Nie├čner: ScanRefer: 3D Object Localization in RGB-D Scans using Natural Language. 16th European Conference on Computer Vision (ECCV), 2020
ScanRefer Baseline0.6422 80.4196 80.3090 80.1832 80.3837 80.2362 8
pairwisemethod0.5779 90.3603 90.2792 90.1746 90.3462 90.2163 9