Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary Segmentation

Benchmark Model Rank Results
unsupervised-semantic-segmentation-with-language-image-pre-training-on-cityscapes-valTrident#2mIoU: 47.6
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-objectTrident#3mIoU: 42.2
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-context-59Trident#3mIoU: 44.3
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-context-60Trident#3mIoU: 40.1
unsupervised-semantic-segmentation-with-language-image-pre-training-on-ade20kTrident#4Mean IoU (val): 26.7
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-vocTrident#4mIoU: 70.8
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-stuff-171Trident#5mIoU: 28.6
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascalvoc-20Trident#5mIoU: 88.7