| SimpleBaseline(DeepLabv3+ with ImageNet pretrained Xception65, single scale inference) | 77.8% | A Simple Baseline for Semi-supervised Semantic Segmentation with Strong Data Augmentation | - |
| U2PL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K, AEL) | 78.51% | Semi-Supervised Semantic Segmentation Using Unreliable Pseudo-Labels | - |
| S4MC | 79.52% | Semi-Supervised Semantic Segmentation via Marginal Contextual Information | - |
| ReCo (DeepLab v3+ with ResNet-101 backbone, ImageNet pretrained) | 68.50% | Bootstrapping Semantic Segmentation with Regional Contrast | - |
| CPS (DeepLab v3+ with ImageNet-pretrained ResNet-101, single scale inference) | 79.21% | Semi-Supervised Semantic Segmentation with Cross Pseudo Supervision | - |
| UniMatch V2 (DINOv2-B) | 84.5% | UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation | - |
| PS-MT (DeepLab v3+ with ImageNet-pretrained ResNet-50, single scale inference) | 78.38% | Perturbed and Strict Mean Teachers for Semi-supervised Semantic Segmentation | - |
| PrevMatch (ResNet-50) | 78.8% | Revisiting and Maximizing Temporal Knowledge in Semi-supervised Semantic Segmentation | - |
| LaserMix (DeepLab v3+, ImageNet pre- trained ResNet50, single scale inference) | 78.3% | LaserMix for Semi-Supervised LiDAR Semantic Segmentation | - |
| Error Localization Network (DeeplabV3 with ResNet-50) | 73.52% | Semi-supervised Semantic Segmentation with Error Localization Network | - |
| GuidedMix-Net(DeepLab v2 with ResNet101, ImageNet pretrained) | 67.5% | GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference | - |
| SemiVL (ViT-B/16) | 80.3% | SemiVL: Semi-Supervised Semantic Segmentation with Vision-Language Guidance | - |
| Adversarial (DeepLab v2 ImageNet pre-trained) | 60.5% | Adversarial Learning for Semi-Supervised Semantic Segmentation | - |
| CW-BASS (DeepLab v3+ with ResNet-50) | 78.43% | CW-BASS: Confidence-Weighted Boundary-Aware Learning for Semi-Supervised Semantic Segmentation | - |
| Dual Teacher (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K) | 79.46 | Switching Temporary Teachers for Semi-Supervised Semantic Segmentation | |
| GIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained) | 65.14% | The GIST and RIST of Iterative Self-Training for Semi-Supervised Segmentation | - |
| ReCo (DeepLab v2 with ResNet-101 backbone, ImageNet pretrained) | 67.53% | Bootstrapping Semantic Segmentation with Regional Contrast | - |
| SegSDE (MTL decoder with ResNet101, ImageNet pretrained, unlabeled image sequences) | 69.38% | Three Ways to Improve Semantic Segmentation with Self-Supervised Depth Estimation | - |
| CorrMatch (Deeplabv3+ with ResNet-101) | 79.4% | CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentation | - |
| CutMix (DeepLab v2, ImageNet pre-trained) | 63.87% | Semi-supervised semantic segmentation needs strong, varied perturbations | - |