Interactive Image Segmentation on DAVIS (NoC Thresholds 85, 90, 95)
4.54NoC @ 90% IoUSAM-REF
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SAM-REFTrain data=COCO+LVIS, Backbone=ViT-H*, SPC/s=0.511, Fusion Strategy=Late Fusion2024.08 | 4.54 | — | 9.1 | |
| AdaptiveClickTrain data=COCO+LVIS, Backbone=VIT-H, Fusion Strategy=Early Fusion2024.08 | 4.6 | — | — | |
| SAM-REFTrain data=COCO+LVIS, Backbone=ViT-B*, SPC/s=0.278, Fusion Strategy=Late Fusion2024.08 | 4.75 | — | 10.45 | |
| FocSAMTrain data=COCO+LVIS, Backbone=VIT-H*, SPC/s=0.467, Fusion Strategy=Late Fusion2024.08 | 4.76 | — | 10.77 | |
| SimpleClickTrain data=COCO+LVIS, Backbone=VIT-H, SPC/s=8.24, Fusion Strategy=Early Fusion2024.08 | 4.78 | — | 8.88 | |
| SAM-REFTrain data=HQSeg-44K, Backbone=ViT-H*, SPC/s=0.511, Fusion Strategy=Late Fusion2024.08 | 4.78 | — | 9.38 | |
| FocalClickTrain data=COCO+LVIS, Backbone=SegFB3-S2384, SPC/s=0.232, Fusion Strategy=Early Fusion2024.08 | 4.9 | — | 10.4 | |
| HQ-SAMTrain data=HQSeg-44K, Backbone=ViT-H*, SPC/s=0.443, Fusion Strategy=Late Fusion2024.08 | 5.06 | — | 9.65 | |
| PseudoClickTrain data=COCO+LVIS, Backbone=HRNet32, Fusion Strategy=Early Fusion2024.08 | 5.11 | — | — | |
| SAMTrain data=SA-1B, Backbone=ViT-H, SPC/s=0.413, Fusion Strategy=Late Fusion2024.08 | 5.19 | — | 10 | |
| InterFormerTrain data=COCO+LVIS, Backbone=VIT-L, SPC/s=0.271, Fusion Strategy=Late Fusion2024.08 | 5.21 | — | — | |
| MFPBackbone=ViT-B2024.04 | 5.32 | 3.92 | 11.27 | |
| RITMTrain data=COCO+LVIS, Backbone=HRNet32, Fusion Strategy=Early Fusion2024.08 | 5.34 | — | 11.52 | |
| SimpleClickBackbone=ViT-B2024.04 | 5.48 | 4.1 | 12.24 | |
| iCMFormerBackbone=ViT-B2024.04 | 5.58 | 4.05 | — | |
| GPCISBackbone=ResNet-502024.04 | 5.89 | 4.37 | 12.42 | |
| InterFormerTrain data=COCO+LVIS, Backbone=VIT-B, SPC/s=0.153, Fusion Strategy=Late Fusion2024.08 | 6.19 | — | — | |
| FocusCutBackbone=ResNet-1012024.04 | 6.22 | 4.85 | — | |
| PseudoClickBackbone=HRNet-182024.04 | 6.57 | 4.81 | — | |
| RITMBackbone=HRNet-182024.04 | 6.71 | 4.94 | 13.87 | |
| EMC-ClickBackbone=HRNet-182024.04 | 6.71 | 5.05 | — | |
| CDNetBackbone=ResNet-342024.04 | 6.89 | 5 | 14.24 | |
| CDNetTrain data=COCO+LVIS, Backbone=ResNet, Fusion Strategy=Early Fusion2024.08 | 7.04 | — | — | |
| FocalClickBackbone=SegF-BO2024.04 | 7.06 | 5.02 | — | |
| f-BRSBackbone=ResNet-1012024.04 | 7.41 | 5.04 | — | |
| F-BRSTrain data=COCO+LVIS, Backbone=ResNet, Fusion Strategy=Early Fusion2024.08 | 7.41 | — | — | |
| BRSBackbone=DenseNet2024.04 | 8.24 | 5.58 | — |