Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text Pairs

Benchmark Model Rank Results
semantic-segmentation-on-cc3m-tagmaskTCL#2mIoU: 60.4
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-context-59TCL#7mIoU: 33.9
unsupervised-semantic-segmentation-with-language-image-pre-training-on-ade20kTCL#8Mean IoU (val): 17.1
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-objectTCL#8mIoU: 31.6
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascal-vocTCL#8mIoU: 55.0
unsupervised-semantic-segmentation-with-language-image-pre-training-on-pascalvoc-20TCL#8mIoU: 83.2
unsupervised-semantic-segmentation-with-language-image-pre-training-on-cityscapes-valTCL#9mIoU: 24.0
unsupervised-semantic-segmentation-with-language-image-pre-training-on-coco-stuff-171TCL#10mIoU: 22.4
open-vocabulary-semantic-segmentation-on-pascalvoc-20TCL#16mIoU: 83.2
open-vocabulary-semantic-segmentation-on-pascal-context-59TCL#24mIoU: 33.9