The GIST and RIST of Iterative Self-Training for Semi-Supervised Segmentation

Benchmark Model Rank Results
semi-supervised-semantic-segmentation-on-cityscapes-100-samples-labeledGIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained)–Validation mIoU: 58.70%
semi-supervised-semantic-segmentation-on-cityscapes-12-5-labeledGIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained)–Validation mIoU: 62.57%
semi-supervised-semantic-segmentation-on-cityscapes-25-labeledGIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained)–Validation mIoU: 65.14%
semi-supervised-semantic-segmentation-on-pascal-voc-2012-12-5-labeledGIST and RIST–Validation mIoU: 70.76%
semi-supervised-semantic-segmentation-on-pascal-voc-2012-2-labeledGIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained)–Validation mIoU: 67.21%
semi-supervised-semantic-segmentation-on-pascal-voc-2012-5-labeledGIST and RIST (DeepLabv2 with ResNet101, MSCOCO pre-trained)–Validation mIoU: 69.40%