Dichotomous Image Segmentation on DIS-TE2 500 (test)
92FmaxPDFNet
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| PDFNetBackbone=Swin-B, Params=94M2025.03 | 92 | 88.6 | 94.8 | 92.4 | 2.7 | |
| DiffDISBackbone=SD-Turbo, Params=865M+2025.03 | 91.7 | 89.5 | 95.1 | 91.3 | 2.6 | |
| MVANetBackbone=Swin-B, Params=93M2025.03 | 91.6 | 87.4 | 94.4 | 91.5 | 3 | |
| MVANet*Backbone=Swin-B, Params=93M, Input=RGB-depth concatenation2025.03 | 90.8 | 86.9 | 93.7 | 91.3 | 3.2 | |
| BiRefNetBackbone=Swin-L, Params=215M2025.03 | 90.6 | 85.7 | 93 | 90 | 3.6 | |
| InSPyReNetBackbone=Swin-B, Params=88M2025.03 | 89.4 | 84.6 | 91.6 | 90.5 | 3.6 | |
| CPNetBackbone=Swin-B & Swin-B, Params=216M2025.03 | 89.2 | 85.5 | 92.9 | 89.9 | 3.6 | |
| DACODBackbone=ResNet-50 & Swin-B, Params=189M2025.03 | 88.4 | 83.9 | 92.5 | 89.1 | 3.9 | |
| GenPerceptBackbone=SD U-Net, Params=865M+2025.03 | 88 | 85.9 | 93.8 | 89.2 | 3.4 | |
| MAGNetBackbone=SMT-T & MbNet-V2, Params=16M2025.03 | 87.6 | 83.3 | 92.3 | 88.6 | 4.1 | |
| RISNetBackbone=PVT-V2-B2, Params=25M2025.03 | 86.9 | 82.8 | 91.8 | 88 | 4.4 | |
| UDUNBackbone=ResNet-50, Params=25M2025.03 | 82.9 | 76.8 | 88.6 | 84.3 | 5.8 | |
| FP-DIS2025.03 | 82.7 | 76.7 | 89.3 | 84.5 | 5.9 | |
| IS-NetBackbone=U2-Net, Params=44M2025.03 | 79.9 | 72.8 | 85.8 | 82.3 | 7 |