Open-Vocabulary Panoptic Segmentation with Text-to-Image Diffusion Models

Benchmark Model Rank Results
open-vocabulary-panoptic-segmentation-on-ade20kODISE(Caption)#7PQ: 23.4
open-vocabulary-panoptic-segmentation-on-ade20kODISE (Label)#8PQ: 22.6
open-vocabulary-semantic-segmentation-on-1ODISE#13mIoU: 57.3
open-vocabulary-semantic-segmentation-on-2ODISE#14mIoU: 29.9
open-vocabulary-semantic-segmentation-on-3ODISE#15mIoU: 11.1
open-vocabulary-semantic-segmentation-on-5ODISE#13mIoU: 84.6
open-vocabulary-semantic-segmentation-on-7ODISE#12mIoU: 14.5
zero-shot-segmentation-on-segmentation-in-theodise#6Mean AP: 38.7