| SemiVL: Semi-Supervised Semantic Segmentation with Vision-Language Guidance | ✓ Link | 76.2 | SemiVL (ViT-B/16) | 2023-11-27 |
| Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation | ✓ Link | 73.0 | UniMatch (DeepLab v3+ with ResNet-101) | 2022-08-21 |
| Confidence-Weighted Boundary-Aware Learning for Semi-Supervised Semantic Segmentation | ✓ Link | 65.87 | CW-BASS (DeepLab v3+ with ResNet-50) | 2025-02-21 |
| Semi-Supervised Semantic Segmentation with Pixel-Level Contrastive Learning from a Class-wise Memory Bank | ✓ Link | 64.9% | SemiSegContrast (DeepLab v3+ with ResNet-50 backbone, MSCOCO pretrained) | 2021-04-27 |
| Three Ways to Improve Semantic Segmentation with Self-Supervised Depth Estimation | ✓ Link | 62.09% | SegSDE (MTL decoder with ResNet101, ImageNet pretrained, unlabeled image sequences) | 2020-12-19 |
| Bootstrapping Semantic Segmentation with Regional Contrast | ✓ Link | 60.28% | ReCo (DeepLab v3+ with ResNet-101 backbone, ImageNet pretrained) | 2021-04-09 |
| Semi-Supervised Semantic Segmentation with Pixel-Level Contrastive Learning from a Class-wise Memory Bank | ✓ Link | 59.4% | SemiSegContrast (DeepLab v2 with ResNet-101 backbone, MSCOCO pretrained) | 2021-04-27 |
| The GIST and RIST of Iterative Self-Training for Semi-Supervised Segmentation | | 58.70% | GIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained) | 2021-03-31 |
| GuidedMix-Net: Learning to Improve Pseudo Masks Using Labeled Images as Reference | ✓ Link | 56.9% | GuidedMix-Net(DeepLab v2 with ResNet101, ImageNet pretrained) | 2021-06-29 |
| Bootstrapping Semantic Segmentation with Regional Contrast | ✓ Link | 56.53% | ReCo (DeepLab v2 with ResNet-101 backbone, ImageNet pretrained) | 2021-04-09 |
| DMT: Dynamic Mutual Training for Semi-Supervised Learning | ✓ Link | 54.80% | DMT (DeepLab v2 MSCOCO/ImageNet pre-trained) | 2020-04-18 |
| ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learning | ✓ Link | 54.07% | ClassMix (DeepLab v2 MSCOCO pretrained) | 2020-07-15 |
| Semi-supervised semantic segmentation needs strong, varied perturbations | ✓ Link | 51.2 | CutMix (DeepLab v2, ImageNet pre-trained) | 2019-06-05 |