List your company

Video Object Segmentation on YouTube-VOS 2018 (val)

86.1J Score (Seen)ISVOS
81.9483.0284.185.18Mar 18, 2020Dec 17, 2020Sep 18, 2021Jun 20, 2022Mar 22, 2023Dec 22, 2023Sep 22, 2024

Evaluation results

MethodLinks
ISVOSw/ BL30K=true2022.1286.186.790.88189———
ISVOSw/ BL30K=false2022.1285.586.390.280.588.8———
XMemw/ BL30K=true2022.1285.186.189.880.389.2———
XMemMemory Mechanism=Memory2023.0385.186.189.880.389.2———
XMemMemory-based=true2023.1285.186.189.880.389.2———
R50-AOSTHierarchical Structure=true, Backbone=ResNet-50, n_l=3, AF=true2022.0385.1—90.278.987.3—85.46.3
SwinB-AOSTHierarchical Structure=true, Backbone=Swin-Base, n_l=3, AF=true2022.0385.1—90.179.688.2—85.85.1
DeAOTw/ BL30K=false2022.1284.98689.980.488.7———
DeAOTMethod Category=Memory-based2023.1284.98689.980.488.7———
DeAOTHierarchical Structure=true2022.0384.9—89.980.488.7—8622.4
BATMAN2022.0884.7—89.879.287.4—85.3—
XMemw/ BL30K=false2022.1284.685.789.380.288.7———
XMemLabeled data=100%2023.0384.685.789.380.288.7———
XMemMethod Category=Memory-based2023.1284.685.789.380.288.7———
XMemAF=true2022.0384.6—89.380.288.7—85.722.6
AOTw/ BL30K=false2022.1284.585.589.579.688.2———
SwinB-AOSTHierarchical Structure=true, Backbone=Swin-Base, n_l=32022.0384.5—89.678.687—84.99.3
SwinB-AOSTHierarchical Structure=true, Backbone=Swin-Base, n_l=3, alpha=infinity during training2022.0384.5—89.47987.5—85.19.3
XMem*Labeled data=100%2023.0384.485.589.18088.3———
SwinB-AOT-LBackbone=Swin-B, Scale=Large2021.0684.3—89.377.986.4—84.59.3
SwinB-AOT-LBackbone=Swin-B2021.0684.3—89.377.986.4—84.5—
SwinB-AOT-Lstatic_image_datasets=true, fps=9.32022.0784.384.589.377.986.44.65——
R50-AOSTHierarchical Structure=true, Backbone=ResNet-50, n_l=3, alpha=infinity during training2022.0384.1—88.978.887.3—84.814.9
UniRef++-LMethod Category=Non-memory, Backbone=Large2023.1283.883.288.576.883.8———
R50-AOSTHierarchical Structure=true, Backbone=ResNet-50, n_l=32022.0383.8—88.879.387.9—8514.9
R50-AOT-LBackbone=ResNet-50, Scale=Large2021.0683.7—88.578.186.1—84.114.9
R50-AOT-LBackbone=ResNet-502021.0683.7—88.578.186.1—84.1—
R50-AOTRegional Segmentation=false2022.0783.7—88.578.186.1—84.114.9
AOT2022.0883.7—88.578.186.1—84.1—
R50-AOT-LLabeled data=100%2023.0383.784.188.578.186.1———
AOT-LStage2 only training=true2024.0983.784.178.188.586.1———
XMemStage2 only training=true2024.0983.784.478.288.587.2———
JIMDStage2 only training=true2024.0983.784.879.188.787.6———
2-shot XMem w/ OursLabeled data=7.3%2023.0383.684.888.579.287.7———
AOSTHierarchical Structure=true, n_l=32022.0383.5—88.578.487—84.415.9
RPCMmulti-scale testing=true2021.1283.3—87.978.986.9—84.3—
STCNExtra Pre-training Data=BL30K2021.0683.284.387.97987.3———
STCN2021.0783.284.387.97987.3——116.8
Adaptive Object Calibration Networkvariant=BaseMS, multi-scale and flip testing=true, external data=true2022.0783.2—87.879.387.3—84.4—
STCNRegional Segmentation=false2022.0783.2—87.97987.3—84.312
STCNw/ BL30K=true2022.1283.284.387.97987.3———
UniRef-LMethod Category=Non-memory, Backbone=Large2023.1283.282.687.576.283.7———
STCN2024.0983.284.37987.987.3———
RPCM2021.1283.1—87.778.586.7—84—
RAVOSRegional Segmentation=true2022.0783.1—87.879.187.4—84.423
RPCMVOS2022.0883.1—87.778.586.7—84—
RPCMw/ BL30K=false2022.1283.18487.778.586.7———
RPCMAF=false2022.0383.1—87.778.586.7—84—
PCVOSvariant=L52022.088384.68879.687.9———
PCVOSAF=true2022.0383—8879.687.9—84.610.5
AOT-LargeScale=Large2021.0682.9—87.977.786.5—83.816
AOT-L2021.0682.9—87.977.786.5—83.8—
AOT-LMethod Category=Memory-based, Variant=Large2023.1282.983.887.977.786.5———
AOT-LHierarchical Structure=true2022.0382.9—87.977.786.5—83.816
CFBI+time per frame (s)=2.17, multi-scale and flip strategy=true2020.1082.883.387.577.385.7———
CFBI+multi-scale and flip testing=true, external data=true2022.0782.8—87.577.385.7—83.3—
Adaptive Object Calibration Networkvariant=MF, multiple historical frames=true, external data=true2022.0782.7—87.478.887.1—84—
QDMNmulti-scale=true2022.0782.7—87.578.486.483.8——
AOT-BaseScale=Base2021.0682.6—87.577.786—83.520.5
AOT-B2021.0682.6—87.577.786—83.5—
AOT-B2022.0782.6—87.477.385.6—83.2—
Adaptive Object Calibration Networkvariant=Base2022.0782.6—87.278.386.3—83.6—
DMN-AOAmultiple historical frames=true, external data=true2022.0782.6—8776.784.8—82.7—
AOT-BMethod Category=Memory-based, Variant=Base2023.1282.683.587.577.786———
AOT-BHierarchical Structure=true2022.0382.6—87.577.786—83.520.5
AOT-Lmultiple historical frames=true, external data=true2022.0782.5—87.577.986.7—83.7—
DMN-AOA2022.0782.5—86.976.284.282.5——
AOT-L2022.0782.5—87.577.986.783.7——
DMN-AOARegional Segmentation=true2022.0782.5—86.976.284.2—82.5—
DMN-AOA2022.0882.582.586.976.282.5———
DMNinference_speed_gt_20fps=false2022.0882.582.586.976.284.2———
AOTinference_speed_gt_20fps=false2022.0882.583.787.577.986.7———
SWEMinference_speed_gt_20fps=true2022.0882.482.886.977.185———
SWEMMemory-based=true2023.1282.482.886.977.185———
UniRef++-R50Method Category=Non-memory, Backbone=ResNet-502023.1282.381.986.875.982.6———
CFBImulti-scale and flip strategy=true2020.0382.282.786.876.985———
LCM2021.0482.28286.775.783.4———
CFBItime per frame (s)=2.51, multi-scale and flip strategy=true2020.1082.282.786.876.985———
CFBIMSsynthetic data for offline training=false, online fine-tuning=false2021.0882.282.786.876.985———
LCMContemporary work=true2021.0682.28286.775.783.4———
LCMuse_static_images=true2021.1082.28286.775.783.4———
LCMexternal training datasets=true2021.1282.2—86.775.783.4—82—
CFBImulti-scale testing=true, flip testing=true2021.1282.2—86.876.985—82.7—
LCMall-frames=true, multiple historical frames=true, external data=true2022.0782.2—86.775.783.4—82—
CFBImulti-scale and flip testing=true, external data=true2022.0782.2—86.876.985—82.7—
LCM2022.0782.2—86.775.783.482——
LCMRegional Segmentation=false2022.0782.2—86.775.783.4—82—
LCM2022.0882.28286.775.783.4———
LCMinference_speed_gt_20fps=false2022.0882.28286.775.783.4———
LCM2022.0882.2—86.775.783.4—82—
LCMstatic_image_datasets=true2022.0782.28286.775.783.44.9——
HMMN2021.0982.182.68776.884.6———
RMNetContemporary work=true2021.0682.181.585.775.782.4———
RMNetSupervision=Fully-supervised2021.0782.181.585.775.782.44.9——
RMNexternal training datasets=true2021.1282.1—85.775.782.4—81.5—
HMMNexternal training datasets=true2021.1282.1—8776.884.6—82.6—
RMNet2022.0782.1—85.775.782.481.5——
HMMN2022.0782.1—8776.884.682.6——
HMMNRegional Segmentation=false2022.0782.1—8776.884.6—82.6—
RMNetRegional Segmentation=true2022.0782.1—85.775.782.4—81.5—