| OneFormer: One Transformer to Rule Universal Image Segmentation | ✓ Link | 68.0 | OneFormer (ConvNeXt-L, single-scale, Mapillary Vistas-Pretrained) | 2022-11-10 |
| Scaling Wide Residual Networks for Panoptic Segmentation | | 67.8 | Panoptic-DeepLab (SWideRNet [1, 1, 4.5], Mapillary, multi-scale) | 2020-11-23 |
| EfficientPS: Efficient Panoptic Segmentation | ✓ Link | 67.1 | EfficientPS | 2020-04-05 |
| Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation | ✓ Link | 66.6 | Axial-DeepLab-XL (Mapillary Vistas, multi-scale) | 2020-03-17 |
| kMaX-DeepLab: k-means Mask Transformer | ✓ Link | 66.2 | kMaX-DeepLab (single-scale) | 2022-07-08 |
| Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation | ✓ Link | 65.5 | Panoptic-Deeplab | 2019-11-22 |
| EfficientPS: Efficient Panoptic Segmentation | ✓ Link | 62.9 | EfficientPS (Cityscapes-fine) | 2020-04-05 |
| SOGNet: Scene Overlap Graph Network for Panoptic Segmentation | ✓ Link | 60 | SOGNet (ResNet-50) | 2019-11-18 |
| Combinatorial Optimization for Panoptic Segmentation: A Fully Differentiable Approach | ✓ Link | 60 | COPS (ResNet-50) | 2021-06-06 |
| Pixelwise Instance Segmentation with a Dynamically Instantiated Network | ✓ Link | 55.4 | Dynamically Instantiated Network | 2017-04-07 |