Unsupervised Video Object Segmentation on DAVIS 2016 (val)
89.6F MeanGSA-Net
Evaluation Results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GSA-NetBackbone=MiT-b2, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=4.1, test time augmentation=true2023.03 | 89.6 | — | — | — | — | — | — | — | — | — | 88.3 | — | — | 88.9 | — | |
| SimulFlowBackbone=MiT-b2, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=25.22024.07 | 89.5 | — | — | — | — | — | — | — | — | — | — | — | — | 88.3 | 87.1 | |
| HFANPublication=ECCV'22, Backbone=MiT-b2, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=2.9, test time augmentation=true2023.03 | 89.3 | — | — | — | — | — | — | — | — | — | 88 | — | — | 88.7 | — | |
| GSA-NetBackbone=ResNet101, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=4.5, test time augmentation=true2023.03 | 89 | — | — | — | — | — | — | — | — | — | 87.9 | — | — | 88.4 | — | |
| GSA-NetBackbone=MiT-b2, Resolution=512x512, Optical Flow=false, Post-processing=false, fps=38.22024.07 | 89 | — | — | — | — | — | — | — | — | — | — | — | — | 88.2 | 87.4 | |
| FakeFlowBackbone=MiT-b2, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=29.52024.07 | 89 | — | — | — | — | — | — | — | — | — | — | — | — | 88.5 | 88 | |
| GSA-NetBackbone=MiT-b2, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=312023.09 | 89 | — | — | — | — | — | — | — | — | — | — | — | — | 88.2 | 87.4 | |
| GFAResolution=512x512, Optical Flow=false, Post-processing=false2024.07 | 88.9 | — | — | — | — | — | — | — | — | — | — | — | — | 88.2 | 87.4 | |
| GFAResolution=512x512, Optical Flow=true, Post-processing=false2023.09 | 88.9 | — | — | — | — | — | — | — | — | — | — | — | — | 88.2 | 87.4 | |
| GSA-NetBackbone=ResNet101, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=41.52023.03 | 88.4 | — | — | — | — | — | — | — | — | — | 87 | — | — | 87.7 | — | |
| DPABackbone=VGG-16, Resolution=512x512, Optical Flow=false, Post-processing=false, fps=19.52024.07 | 88.4 | — | — | — | — | — | — | — | — | — | — | — | — | 87.6 | 86.8 | |
| DPABackbone=VGG-16, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=19.52023.09 | 88.4 | — | — | — | — | — | — | — | — | — | — | — | — | 87.6 | 86.8 | |
| TMO++Backbone=MiT-b1, Resolution=Multiple, Optical Flow=false, Post-processing=true, fps=4.3, Test-time augmentation=true2023.09 | 88.3 | — | — | — | — | — | — | — | — | — | — | — | — | 88.2 | 88 | |
| HFANPublication=ECCV'22, Backbone=MiT-b2, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=18.42023.03 | 88.2 | — | — | — | — | — | — | — | — | — | 86.8 | — | — | 87.5 | — | |
| HFANBackbone=MiT-b2, Resolution=512x512, Optical Flow=false, Post-processing=false, fps=12.82024.07 | 88.2 | — | — | — | — | — | — | — | — | — | — | — | — | 87.5 | 86.8 | |
| HFANPublication=ECCV'22, Backbone=ResNet101, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=2.5, test time augmentation=true2023.03 | 87.9 | — | — | — | — | — | — | — | — | — | 87.3 | — | — | 87.6 | — | |
| DATTTBackbone=MiT-b1, Resolution=512x512, Optical Flow=false, Post-processing=false2023.09 | 87.9 | — | — | — | — | — | — | — | — | — | — | — | — | 87 | 86 | |
| TMOBackbone=MiT-b1, Resolution=384x384, Optical Flow=false, Post-processing=false, fps=66.72023.09 | 87.8 | — | — | — | — | — | — | — | — | — | — | — | — | 87.2 | 86.5 | |
| TMO++Backbone=MiT-b1, Resolution=384x384, Optical Flow=false, Post-processing=true, fps=54.32023.09 | 87.8 | — | — | — | — | — | — | — | — | — | — | — | — | 87.2 | 86.5 | |
| OASTPublication=ICCV'23, Backbone=MiT-b2, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false2023.03 | 87.4 | — | — | — | — | — | — | — | — | — | 86.6 | — | — | 87 | — | |
| GSA-NetBackbone=MiT-b2, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=38.22023.03 | 87.4 | — | — | — | — | — | — | — | — | — | 87.4 | — | — | 88.2 | — | |
| OASTBackbone=MobileViT3D, Resolution=384x640, Optical Flow=false, Post-processing=false2024.07 | 87.4 | — | — | — | — | — | — | — | — | — | — | — | — | 87 | 86.6 | |
| SimulFlowBackbone=MiT-b1, Resolution=384x384, Optical Flow=true, Post-processing=false, fps=53.22023.09 | 87.4 | — | — | — | — | — | — | — | — | — | — | — | — | 86.7 | 86.3 | |
| HFANPublication=ECCV'22, Backbone=ResNet101, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=192023.03 | 87.3 | — | — | — | — | — | — | — | — | — | 86.6 | — | — | 87 | — | |
| HFANBackbone=MiT-b1, Resolution=512x512, Optical Flow=false, Post-processing=false, fps=18.42023.09 | 87.1 | — | — | — | — | — | — | — | — | — | — | — | — | 86.7 | 86.2 | |
| TMO++Backbone=ResNet-101, Resolution=Multiple, Optical Flow=false, Post-processing=true, fps=2.3, Test-time augmentation=true2023.09 | 87 | — | — | — | — | — | — | — | — | — | — | — | — | 87.2 | 87.4 | |
| Ours2021.12 | 86.7 | 84.5 | — | — | — | 92.7 | 2.8 | 93.3 | 0.8 | 16.3 | — | — | — | — | — | |
| IMPPublication=AAAI'22, Backbone=ResNet50, Optical Flow (OF)=false, Post-Processing (PP)=false, FPS=1.792023.03 | 86.7 | — | — | — | — | — | — | — | — | — | 84.5 | — | — | 85.6 | — | |
| IMPBackbone=ResNet-50, Optical Flow=false, Post-processing=false, fps=1.792024.07 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | 85.6 | 84.5 | |
| IMPBackbone=ResNet-50, Optical Flow=false, Post-processing=false, fps=1.792023.09 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | 85.6 | 84.5 | |
| TMOPublication=WACV'23, Backbone=ResNet101, Resolution=384x384, Optical Flow (OF)=true, Post-Processing (PP)=true, FPS=43.22023.03 | 86.6 | — | — | — | — | — | — | — | — | — | 85.6 | — | — | 86.1 | — | |
| TMOBackbone=ResNet-101, Resolution=384x384, Optical Flow=false, Post-processing=false, fps=43.22024.07 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | 86.1 | 85.6 | |
| TMOBackbone=ResNet-101, Resolution=384x384, Optical Flow=false, Post-processing=false, fps=43.22023.09 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | 86.1 | 85.6 | |
| TMO++Backbone=ResNet-101, Resolution=384x384, Optical Flow=false, Post-processing=true, fps=36.52023.09 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | 86.1 | 85.6 | |
| D2Conv3DBackbone=ir-CSN-152, Resolution=480x854, Optical Flow=false, Post-processing=false2024.07 | 86.5 | — | — | — | — | — | — | — | — | — | — | — | — | 86 | 85.5 | |
| D2Conv3DBackbone=ir-CSN-152, Resolution=480x854, Optical Flow=false, Post-processing=false2023.09 | 86.5 | — | — | — | — | — | — | — | — | — | — | — | — | 86 | 85.5 | |
| PMNPublication=WACV'23, Backbone=VGG16, Resolution=352x352, Optical Flow (OF)=true, Post-Processing (PP)=false2023.03 | 86.4 | — | — | — | — | — | — | — | — | — | 85.4 | — | — | 85.9 | — | |
| PMNBackbone=VGG-16, Resolution=352x352, Optical Flow=false, Post-processing=false, fps=41.32024.07 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | 85.9 | 85.4 | |
| PMNBackbone=VGG-16, Resolution=352x352, Optical Flow=true, Post-processing=false, fps=41.32023.09 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | 85.9 | 85.4 | |
| OASTBackbone=3D-ResNet-101, Resolution=384x640, Optical Flow=false, Post-processing=false2023.09 | 86.3 | — | — | — | — | — | — | — | — | — | — | — | — | 85.9 | 85.4 | |
| HCPNBackbone=ResNet-101, Resolution=512x512, Optical Flow=true, Post-processing=true, fps=22023.09 | 85.4 | — | — | — | — | — | — | — | — | — | — | — | — | 85.6 | 85.8 | |
| TransportNetPublication=ICCV'21, Backbone=ResNet101, Resolution=512x512, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=12.52023.03 | 85 | — | — | — | — | — | — | — | — | — | 84.5 | — | — | 84.8 | — | |
| TransportNetBackbone=ResNet-101, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=12.52024.07 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | 84.8 | 84.5 | |
| TransportNetBackbone=ResNet-101, Resolution=512x512, Optical Flow=true, Post-processing=false, fps=12.52023.09 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | 84.8 | 84.5 | |
| 3DC-Seg (Dense)Optical Flow=false, CRF=false, Multi-Scale=false, Time (s/frame)=0.842020.08 | 84.7 | 84.3 | — | — | 84.5 | — | — | — | — | — | — | — | — | — | — | |
| RTNetPublication=CVPR'21, Backbone=ResNet101, Resolution=384x672, Optical Flow (OF)=true, Post-Processing (PP)=true2023.03 | 84.7 | — | — | — | — | — | — | — | — | — | 85.6 | — | — | 85.2 | — | |
| RTNetBackbone=ResNet-101, Resolution=384x672, Optical Flow=true, Post-processing=true2024.07 | 84.7 | — | — | — | — | — | — | — | — | — | — | — | — | 85.2 | 85.6 | |
| RTNetBackbone=ResNet-101, Resolution=384x672, Optical Flow=true, Post-processing=true2023.09 | 84.7 | — | — | — | — | — | — | — | — | — | — | — | — | 85.2 | 85.6 | |
| AMC-NetPublication=ICCV'21, Backbone=ResNet101, Resolution=384x384, Optical Flow (OF)=true, Post-Processing (PP)=true, FPS=17.52023.03 | 84.6 | — | — | — | — | — | — | — | — | — | 84.5 | — | — | 84.6 | — | |
| AMC-NetBackbone=ResNet-101, Resolution=384x384, Optical Flow=true, Post-processing=false, fps=17.52024.07 | 84.6 | — | — | — | — | — | — | — | — | — | — | — | — | 84.6 | 84.5 | |
| AMC-NetBackbone=ResNet-101, Resolution=384x384, Optical Flow=true, Post-processing=false, fps=17.52023.09 | 84.6 | — | — | — | — | — | — | — | — | — | — | — | — | 84.6 | 84.5 | |
| F2Net2020.12 | 84.4 | 83.1 | — | — | 83.7 | 95.7 | 0 | 92.3 | 0.8 | 20.9 | — | — | — | — | — | |
| F2Net2021.12 | 84.4 | 83.1 | — | — | — | 95.7 | 0 | 92.3 | 0.8 | 20.9 | — | — | — | — | — | |
| F2NetPublication=AAAI'21, Backbone=DeepLabV3, Resolution=473x473, Optical Flow (OF)=false, Post-Processing (PP)=false, FPS=102023.03 | 84.4 | — | — | — | — | — | — | — | — | — | 83.1 | — | — | 83.7 | — | |
| F2NetBackbone=DeepLabv3, Resolution=473x473, Optical Flow=false, Post-processing=false, fps=102024.07 | 84.4 | — | — | — | — | — | — | — | — | — | — | — | — | 83.7 | 83.1 | |
| 3DC-SegOptical Flow=false, CRF=false, Multi-Scale=false, Time (s/frame)=0.222020.08 | 84.3 | 83.9 | — | — | 84.1 | — | — | — | — | — | — | — | — | — | — | |
| FSNetPublication=ICCV'21, Backbone=ResNet50, Resolution=352x352, Optical Flow (OF)=true, Post-Processing (PP)=true, FPS=12.52023.03 | 83.1 | — | — | — | — | — | — | — | — | — | 83.4 | — | — | 83.3 | — | |
| FSNetBackbone=ResNet-50, Resolution=352x352, Optical Flow=false, Post-processing=true, fps=12.52024.07 | 83.1 | — | — | — | — | — | — | — | — | — | — | — | — | 83.3 | 83.4 | |
| FSNetBackbone=ResNet-50, Resolution=352x352, Optical Flow=true, Post-processing=true, fps=12.52023.09 | 83.1 | — | — | — | — | — | — | — | — | — | — | — | — | 83.3 | 83.4 | |
| DFNet2021.12 | 81.8 | 83.4 | — | — | — | — | — | — | — | 15.9 | — | — | — | — | — | |
| DFNetPublication=ECCV'20, Backbone=DeepLabV3, Optical Flow (OF)=false, Post-Processing (PP)=true, FPS=3.572023.03 | 81.8 | — | — | — | — | — | — | — | — | — | 83.4 | — | — | 82.6 | — | |
| DFNetBackbone=DeepLabv3, Optical Flow=false, Post-processing=false, fps=3.572024.07 | 81.8 | — | — | — | — | — | — | — | — | — | — | — | — | 82.6 | 83.4 | |
| GMVOS2021.12 | 81.2 | 82.5 | — | — | — | 94.3 | 4.2 | 90.3 | 5.6 | 19.8 | — | — | — | — | — | |
| IMCNetBackbone=ResNet-101, Resolution=480x480, Optical Flow=true, Post-processing=false, fps=202023.09 | 81.1 | — | — | — | — | — | — | — | — | — | — | — | — | 81.9 | 82.7 | |
| MAT2020.12 | 80.7 | 82.4 | — | — | 81.5 | 94.5 | 5.5 | 90.2 | 4.5 | 21.6 | — | — | — | — | — | |
| MATNet2021.12 | 80.7 | 82.4 | — | — | — | 94.5 | 5.5 | 90.2 | 4.5 | 21.6 | — | — | — | — | — | |
| MATNetPublication=AAAI'20, Backbone=ResNet101, Resolution=473x473, Optical Flow (OF)=true, Post-Processing (PP)=false, FPS=202023.03 | 80.7 | — | — | — | — | — | — | — | — | — | 82.4 | — | — | 81.6 | — | |
| WCS-NetPublication=ECCV'20, Backbone=EfficientNetV2, Resolution=320x320, Optical Flow (OF)=false, Post-Processing (PP)=false, FPS=33.32023.03 | 80.7 | — | — | — | — | — | — | — | — | — | 82.2 | — | — | 81.5 | — | |
| MATNetBackbone=ResNet-101, Resolution=473x473, Optical Flow=true, Post-processing=true, fps=202024.07 | 80.7 | — | — | — | — | — | — | — | — | — | — | — | — | 81.6 | 82.4 | |
| MATNetBackbone=ResNet-101, Resolution=473x473, Optical Flow=true, Post-processing=false, fps=202023.09 | 80.7 | — | — | — | — | — | — | — | — | — | — | — | — | 81.6 | 82.4 | |
| STEm-SegOptical Flow=false, CRF=false, Multi-Scale=true, Time (s/frame)=1.422020.08 | 80.6 | 80.6 | — | — | 80.6 | — | — | — | — | — | — | — | — | — | — | |
| AnDiff2020.12 | 80.5 | 81.7 | — | — | 81.1 | 90.9 | 2.2 | 85.1 | 0.6 | 21.4 | — | — | — | — | — | |
| AnDiff2021.12 | 80.5 | 81.7 | — | — | — | 90.9 | 2.2 | 85.1 | 0.6 | 21.4 | — | — | — | — | — | |
| AD-Net + Inst-PruningOptical Flow=false, CRF=true, Multi-Scale=false, Post-processing=heuristic, Time (s/frame)=2.942020.08 | 80.5 | 81.7 | — | — | 81.1 | — | — | — | — | — | — | — | — | — | — | |
| COS2020.12 | 79.5 | 80.5 | — | — | 80 | 93.1 | 4.4 | 89.5 | 5 | 18.4 | — | — | — | — | — | |
| COSNet2021.12 | 79.4 | 80.5 | — | — | — | 93.1 | 4.4 | 89.5 | 5 | 18.4 | — | — | — | — | — | |
| AGNN2021.12 | 79.1 | 80.7 | — | — | — | 94 | 0 | 90.5 | 0 | 33.7 | — | — | — | — | — | |
| AGNNOptical Flow=false, CRF=true, Multi-Scale=true, Time (s/frame)=2.962020.08 | 79.1 | 80.7 | — | — | 79.9 | — | — | — | — | — | — | — | — | — | — | |
| FEM-NetBackbone=ResNet-101, Resolution=473x473, Optical Flow=true, Post-processing=false, fps=72023.09 | 78.4 | — | — | — | — | — | — | — | — | — | — | — | — | 79.9 | 76.9 | |
| AD-NetOptical Flow=false, CRF=false, Multi-Scale=false, Time (s/frame)=0.382020.08 | 78.2 | 79.4 | — | — | 78.8 | — | — | — | — | — | — | — | — | — | — | |
| MoA2020.12 | 77.4 | 77.2 | — | — | 77.3 | 87.8 | 5 | 84.4 | 3.3 | 27.9 | — | — | — | — | — | |
| AGS2020.12 | 77.4 | 79.7 | — | — | 78.6 | 91.1 | 1.9 | 85.8 | 1.6 | 26.7 | — | — | — | — | — | |
| AGS2021.12 | 77.4 | 79.7 | — | — | — | 91.1 | 1.9 | 85.8 | 0 | 26.7 | — | — | — | — | — | |
| MotAdaptOptical Flow=false, CRF=false, Multi-Scale=false2020.08 | 77.4 | 77.2 | — | — | 77.3 | — | — | — | — | — | — | — | — | — | — | |
| 3D-CNNOptical Flow=false, CRF=false, Multi-Scale=false, Time (s/frame)=0.382020.08 | 77.2 | 78.3 | — | — | 77.8 | — | — | — | — | — | — | — | — | — | — | |
| LSMO2020.12 | 75.9 | 78.2 | — | — | 77.1 | 89.1 | 4.1 | 84.7 | 3.5 | 21.2 | — | — | — | — | — | |
| EPO2020.12 | 75.5 | 80.6 | — | — | 78.1 | 95.2 | 2.2 | 87.9 | 2.4 | 19.3 | — | — | — | — | — | |
| PDB2020.12 | 74.5 | 77.2 | — | — | 75.9 | 90.1 | 0.9 | 84.4 | -0.2 | 29.1 | — | — | — | — | — | |
| LSMO2021.12 | 74.5 | 78.2 | — | — | — | 87.8 | 4.1 | 84.7 | 3.5 | 21.2 | — | — | — | — | — | |
| PDBOptical Flow=false, CRF=true, Multi-Scale=false2020.08 | 74.5 | 77.2 | — | — | 75.9 | — | — | — | — | — | — | — | — | — | — | |
| LVO2020.12 | 72.1 | 75.9 | — | — | 74 | 89.1 | 0 | 83.4 | 1.3 | 26.5 | — | — | — | — | — | |
| LVO2021.12 | 72.1 | 75.9 | — | — | — | 89.1 | 0 | 83.4 | 1.3 | 66.7 | — | — | — | — | — | |
| PDB2021.12 | 72.1 | 77.2 | — | — | — | 91.1 | 0.9 | 83.5 | -0.2 | 29.1 | — | — | — | — | — | |
| LVOOptical Flow=true, CRF=true, Multi-Scale=false2020.08 | 72.1 | 75.9 | — | — | 74 | — | — | — | — | — | — | — | — | — | — | |
| ARP2020.12 | 70.6 | 76.2 | — | — | 73.4 | 91.1 | 7 | 83.5 | 7.9 | 39.3 | — | — | — | — | — | |
| MotAdapt2021.12 | 70.6 | 77.2 | — | — | — | 93.1 | 5 | 84.4 | 3.3 | 27.9 | — | — | — | — | — | |
| ARPOptical Flow=true, CRF=false, Multi-Scale=false2020.08 | 70.6 | 76.2 | — | — | 73.4 | — | — | — | — | — | — | — | — | — | — | |
| UOVOS2020.12 | 68 | 73.9 | — | — | 70.9 | 88.5 | 0.6 | 80.6 | 0.7 | 39 | — | — | — | — | — | |
| UOVOS2021.12 | 68 | 73.9 | — | — | — | 88.5 | 0.6 | 80.6 | 0.7 | 39 | — | — | — | — | — | |
| SFL2021.12 | 66.7 | 64.7 | — | — | — | 81.4 | 6.2 | 77.1 | 5.1 | 60.3 | — | — | — | — | — |