Instance Segmentation on COCO (AP, AP50, AP75)
37.84APDeCon-ML-L
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DeCon-ML-LPret. Dataset=COCO+, Backbone=ResNet-50, Pret. Dec.=FPN2025.03 | 37.84 | 59.41 | 40.75 | |
| DeCon-SLPret. Dataset=COCO+, Backbone=ResNet-50, Pret. Dec.=FCN2025.03 | 37.75 | 59.4 | 40.48 | |
| DeCon-ML-LPret. Dataset=ImageNet-1K, Backbone=ResNet-50, Pret. Dec.=FPN2025.03 | 37.73 | 59.08 | 40.68 | |
| SlotCon (2022)Pret. Dataset=ImageNet-1K, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 37.59 | 58.97 | 40.49 | |
| SlotCon (2022)Pret. Dataset=COCO+, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 37.57 | 59.07 | 40.45 | |
| DeCon-ML-LPret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=FPN2025.03 | 37.12 | 58.35 | 39.94 | |
| PixCon (2024)Pret. Dataset=COCO+, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 37.1 | — | — | |
| DeCon-ML-SPret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=FPN2025.03 | 36.94 | 58.2 | 39.63 | |
| DeCon-SLPret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=FCN2025.03 | 36.92 | 58.12 | 39.78 | |
| PixCon (2024)Pret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 36.8 | 57.93 | 39.62 | |
| SlotCon (2022)Pret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 36.8 | 57.98 | 39.54 | |
| DINO (2021)Pret. Dataset=ImageNet-1K, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 36.47 | 57.49 | 39.2 | |
| SupervisedPret. Dataset=ImageNet-1K, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 35.9 | 56.6 | 38.6 | |
| SoCo-D (2024)Pret. Dataset=COCO, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 35.1 | 56.9 | 37.6 | |
| Random init.Pret. Dataset=-, Backbone=ResNet-50, Pret. Dec.=-2025.03 | 29.9 | 47.9 | 32 |