| Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation | ✓ Link | 59.5 | - | - | Mask DINO (single scale) | 2022-06-06 |
| kMaX-DeepLab: k-means Mask Transformer | ✓ Link | 58.5 | 49.0 | 64.8 | kMaX-DeepLab (single-scale) | 2022-07-08 |
| Masked-attention Mask Transformer for Universal Image Segmentation | ✓ Link | 58.3 | 48.1 | 65.1 | Mask2Former (Swin-L) | 2021-12-02 |
| Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers | ✓ Link | 56.2 | 47.0 | 62.3 | Panoptic SegFormer (Swin-L) | 2021-09-08 |
| Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers | ✓ Link | 55.8 | 46.5 | 61.9 | Panoptic SegFormer (PVTv2-B5) | 2021-09-08 |
| CMT-DeepLab: Clustering Mask Transformers for Panoptic Segmentation | ✓ Link | 55.7 | 46.8 | 61.6 | CMT-DeepLab (single-scale) | 2022-06-17 |
| K-Net: Towards Unified Image Segmentation | ✓ Link | 55.2 | 46.2 | 61.2 | K-Net (Swin-L) | 2021-06-28 |
| MaskConver: Revisiting Pure Convolution Model for Panoptic Segmentation | ✓ Link | 53.6 | 58.9 | 45.6 | MaskConver (ResNet50, single-scale) | 2023-12-11 |
| Per-Pixel Classification is Not All You Need for Semantic Segmentation | ✓ Link | 53.3 | 44.5 | 59.1 | MaskFormer (Swin-L) | 2021-07-13 |
| Fully Convolutional Networks for Panoptic Segmentation | ✓ Link | 52.7 | | 59.4 | Panoptic FCN* (Swin-L) | 2020-12-01 |
| REFINE: Prediction Fusion Network for Panoptic Segmentation | | 51.5 | 39.2 | 59.6 | REFINE (ResNeXt-101-DCN) | 2020-12-15 |
| MaX-DeepLab: End-to-End Panoptic Segmentation with Mask Transformers | ✓ Link | 51.3 | 42.4 | 57.2 | MaX-DeepLab-L (single-scale) | 2020-12-01 |
| Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers | ✓ Link | 50.9 | 43.0 | 56.2 | Panoptic SegFormer (ResNet-101) | 2021-09-08 |
| Panoptic SegFormer: Delving Deeper into Panoptic Segmentation with Transformers | ✓ Link | 50.2 | 42.4 | 55.3 | Panoptic SegFormer (ResNet-50) | 2021-09-08 |
| DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution | ✓ Link | 50 | 37.2 | 58.5 | DetectoRS (ResNeXt-101-64x4d, multi-scale) | 2020-06-03 |
| REFINE: Prediction Fusion Network for Panoptic Segmentation | | 49.6 | 37.7 | 57.5 | REFINE (ResNet-101-DCN) | 2020-12-15 |
| SpatialFlow: Bridging All Tasks for Panoptic Segmentation | ✓ Link | 48.5 | 37.9 | 55.5 | SpatialFlow(ResNet-101-FPN) | 2019-10-19 |
| Ada-Segment: Automated Multi-loss Adaptation for Panoptic Segmentation | | 48.5 | 37.6 | 55.7 | Ada-Segment (ResNet-101-DCN) | 2020-12-07 |
| K-Net: Towards Unified Image Segmentation | ✓ Link | 48.3 | 39.7 | 54 | K-Net (R101-FPN-DCN) | 2021-06-28 |
| SOGNet: Scene Overlap Graph Network for Panoptic Segmentation | ✓ Link | 47.8 | | | SOGNet (ResNet-101-FPN) | 2019-11-18 |
| Fully Convolutional Networks for Panoptic Segmentation | ✓ Link | 47.5 | 38.2 | 53.7 | Panoptic FCN*++ (DCN-101-FPN) | 2020-12-01 |
| UPSNet: A Unified Panoptic Segmentation Network | ✓ Link | 46.6 | 36.7 | 53.2 | UPSNet (ResNet-101-FPN) | 2019-01-12 |
| Learning Instance Occlusion for Panoptic Segmentation | ✓ Link | 46.6 | 35.7 | 54.0 | OCFusion (ResNeXt-101-FPN) | 2019-06-13 |
| Scaling Wide Residual Networks for Panoptic Segmentation | | 46.5 | 38.2 | 52.0 | Panoptic-DeepLab (SWideRNet-[1, 1, 4], multi-scale) | 2020-11-23 |
| Attention-guided Unified Network for Panoptic Segmentation | | 46.5 | 32.5 | 55.8 | AUNet (ResNext-152-FPN) | 2018-12-10 |
| Attention-guided Unified Network for Panoptic Segmentation | | 45.5 | 31.6 | 54.7 | AUNet (ResNet-152-FPN) | 2018-12-10 |
| Attention-guided Unified Network for Panoptic Segmentation | | 45.2 | 31.3 | 54.4 | AUNet (ResNet-101-FPN) | 2018-12-10 |
| Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation | ✓ Link | 44.2 | 36.8 | 49.2 | Axial-DeepLab-L (multi-scale) | 2020-03-17 |
| Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation | ✓ Link | 43.6 | 35.6 | 48.9 | Axial-DeepLab-L | 2020-03-17 |
| AdaptIS: Adaptive Instance Selection Network | | 42.8 | 31.8 | 50.1 | AdaptIS (ResNeXt-101) | 2019-09-17 |
| Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation | ✓ Link | 41.4 | 35.9 | 45.1 | Panoptic-DeepLab (Xception-71) | 2019-11-22 |
| Panoptic Feature Pyramid Networks | ✓ Link | 40.9 | 29.7 | 48.3 | Panoptic FPN | 2019-01-08 |
| Learning to Fuse Things and Stuff | | 40.7 | 31.0 | 47.0 | TASCNet | 2018-12-04 |
| EPSNet: Efficient Panoptic Segmentation Network with Cross-layer Attention Fusion | ✓ Link | 38.9 | 31.0 | 44.1 | EPSNet (ResNet-101-FPN) | 2020-03-23 |
| Combinatorial Optimization for Panoptic Segmentation: A Fully Differentiable Approach | ✓ Link | 38.5 | 34.8 | 41.0 | COPS (ResNet-50) | 2021-06-06 |
| Pixel Consensus Voting for Panoptic Segmentation | | 37.7 | 33.1 | 40.7 | PCV (ResNet-50) | 2020-04-04 |
| Generator evaluator-selector net for panoptic image segmentation and splitting unfamiliar objects into parts | ✓ Link | 33.7 | 31.5 | 35.1 | GES Net | 2019-08-24 |
| Panoptic Segmentation with a Joint Semantic and Instance Segmentation Network | | 27.2 | 23.4 | 29.6 | JSIS-Net | 2018-09-06 |