Semi Supervised Semantic Segmentation On 7
评估指标
Validation mIoU
评测结果
各个模型在此基准测试上的表现结果
| Paper Title | Repository | ||
|---|---|---|---|
| ReCo (DeepLab v3+ with ResNet-101 backbone, ImageNet pre-trained) | 63.60% | Bootstrapping Semantic Segmentation with Regional Contrast | |
| ReCo (DeepLab v2 with ResNet-101 backbone, ImageNet pre-trained) | 63.16% | Bootstrapping Semantic Segmentation with Regional Contrast | |
| DMT (DeepLab v2 MSCOCO pre-trained) | 63.04% | DMT: Dynamic Mutual Training for Semi-Supervised Learning | |
| CutMix (DeepLab v3+ ImageNet pre-trained) | 59.52% | Semi-supervised semantic segmentation needs strong, varied perturbations | |
| ClassMix (DeepLab v2 MSCOCO pretrained) | 54.18% | ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learning | |
| CutMix (DeepLab v2 ImageNet pre-trained) | 53.79% | Semi-supervised semantic segmentation needs strong, varied perturbations |
0 of 6 row(s) selected.