Results for InsSSM
Submission data
Full name | 3D Scene Instance Segmentation with Superpoint Select and Merge |
Description | SGIFormer: Semantic-guided and Geometric-enhanced Interleaving Transformer for 3D Instance Segmentation |
Input Data Types | Uses Color,Uses Geometry Uses 3D |
Programming language(s) | Python |
Hardware | 4090 |
Submission creation date | 7 Feb, 2024 |
Last edited | 10 Jul, 2024 |
3D semantic instance results
Info | avg ap 50% | bathtub | bed | bookshelf | cabinet | chair | counter | curtain | desk | door | otherfurniture | picture | refrigerator | shower curtain | sink | sofa | table | toilet | window |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0.799 | 1.000 | 0.915 | 0.710 | 0.729 | 0.925 | 0.664 | 0.670 | 0.770 | 0.766 | 0.739 | 0.737 | 0.700 | 1.000 | 0.792 | 0.829 | 0.815 | 0.997 | 0.625 |