Instance Segmentation on KINS (test)
68.5mIoU (%)Swin-T + Mask R-CNN + Our Plugin
Evaluation results
| Method | Links | |
|---|---|---|
| Swin-T + Mask R-CNN + Our PluginBackbone=Swin-T, Input=GT box, Module=Tri-Layer Plugin, Training Data=COCO2022.10 | 68.5 | |
| Yuan et alInput=GT box, Note=Reference [38] results2022.10 | 67.2 | |
| Swin-T + Mask R-CNN + Bi-LayerBackbone=Swin-T, Input=GT box, Module=Bi-Layer, Training Data=COCO2022.10 | 67 | |
| Swin-T + Mask R-CNNBackbone=Swin-T, Input=GT box, Training Data=COCO2022.10 | 66.6 |