TransFGU: A Top-down Approach to Fine-Grained Unsupervised Semantic Segmentation

Benchmark Model Rank Results
unsupervised-semantic-segmentation-on-coco-6TransFGU (ViT-S/8)#2mIoU: 11.93Pixel Accuracy: 34.32
unsupervised-semantic-segmentation-on-coco-8TransFGU (ViT-S/8)#3mIoU: 12.7Pixel Accuracy: 64.3