Panoptic Segmentation on Mapillary val
Metric: mIoU (higher is better)
Metric: mIoU (higher is better)
| # | Model↕ | mIoU▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | OneFormer (DiNAT-L, single-scale) | 61.7 | No | OneFormer: One Transformer to Rule Universal Ima... | 2022-11-10 | Code |
| 2 | OneFormer (ConvNeXt-L, single-scale) | 61.6 | No | OneFormer: One Transformer to Rule Universal Ima... | 2022-11-10 | Code |
| 3 | Panoptic-DeepLab (SWideRNet-(1, 1, 4.5), multi-scale) | 60 | No | Scaling Wide Residual Networks for Panoptic Segm... | 2020-11-23 | - |
| 4 | Axial-DeepLab-L (multi-scale) | 58.4 | No | Axial-DeepLab: Stand-Alone Axial-Attention for P... | 2020-03-17 | Code |
| 5 | AdaptIS (ResNeXt-101) | 56.8 | No | AdaptIS: Adaptive Instance Selection Network | 2019-09-17 | - |