Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation

Open paper
Benchmark Model Rank Results
panoptic-segmentation-on-coco-test-devMask DINO (single scale)#1PQ: 59.5
semantic-segmentation-on-ade20k-valMaskDINO-SwinL#5mIoU: 60.8
instance-segmentation-on-coco-minivalMasK DINO (SwinL, multi-scale)#6mask AP: 54.5
instance-segmentation-on-coco-minivalMask DINO (SwinL)#12mask AP: 52.6
instance-segmentation-on-coco-test-devMasK DINO (SwinL, multi-scale)#6mask AP: 54.7
instance-segmentation-on-coco-test-devMask DINO (SwinL, single -scale)#13mask AP: 52.8
panoptic-segmentation-on-coco-minivalMasK DINO (SwinL,single-scale)#6PQ: 59.4AP: 50.9
semantic-segmentation-on-ade20kMasK DINO (SwinL, multi-scale)#12Validation mIoU: 60.8Params (M): 223

Rank counts only results with a code link.