Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning

Benchmark Model Rank Results
semi-supervised-semantic-segmentation-on-1AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#10Validation mIoU: 79.01%
semi-supervised-semantic-segmentation-on-15AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#4Validation mIoU: 80.29%
semi-supervised-semantic-segmentation-on-2AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#8Validation mIoU: 77.9%
semi-supervised-semantic-segmentation-on-21AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#11Validation mIoU: 77.2
semi-supervised-semantic-segmentation-on-22AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#9Validation mIoU: 75.83%
semi-supervised-semantic-segmentation-on-35AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#1Validation mIoU: 74.28
semi-supervised-semantic-segmentation-on-4AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#13Validation mIoU: 77.57%
semi-supervised-semantic-segmentation-on-41AEL#3Validation mIoU: 28.4
semi-supervised-semantic-segmentation-on-42AEL#3Validation mIoU: 33.2
semi-supervised-semantic-segmentation-on-8AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#6Validation mIoU: 80.28%
semi-supervised-semantic-segmentation-on-9AEL (DeepLab v3+ with ResNet-101 pretraind on ImageNet-1K)#13Validation mIoU: 78.06