List your company
ResearchFormerly WizWand

Optical Flow on Sintel (test)

1.71AEPE (Final)VideoFlow-BOF
1.61162.27582.943.6042Apr 6, 2021Oct 16, 2021Apr 27, 2022Nov 6, 2022May 18, 2023Nov 27, 2023Jun 7, 2024
Updated 8mo ago

Evaluation results

MethodLinks
VideoFlow-BOFTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.111.711
StreamFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.111.871.04
FlowFormer++Training Data=C+T+S+K+H2023.031.941.07
ProMotionTraining=C+T+S+K(+H)2024.062.011.04
TransFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.081.06
FlowFormerTraining Data=C+T+S+K+H2022.032.091.16
FlowFormerTraining Data=C+T+S+K+H2023.032.091.16
FlowFormerTraining=C+T+S+K+H, Params (M)=18.22023.032.091.16
FlowFormerTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.091.16
FlowFormerTraining=C+T+S+K(+H), Tile technique=true2024.062.121.2
SKFlowTraining Data=C+T+S+K+H, warm-start strategy=true2023.032.231.28
SKFlowTraining=C+T+S+K+H, Params (M)=6.32023.032.231.28
SKFlowTraining=C+T+S+K(+H)2024.062.231.28
SKFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=true2023.112.271.28
KPA-FlowTraining=C+T+S+K(+H)2024.062.361.35
CRAFTTraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.032.421.45
CRAFTTraining dataset=Undisclosed, undisclosed method=true2022.042.421.45
CRAFTTraining Data=C+T+S+K+H2023.032.421.45
CRAFTTraining=C+T+S+K+H, Params (M)=6.32023.032.421.45
CRAFTTraining=C+T+S+K(+H)2024.062.421.45
AnyFlow (dynamic)Training=C+T+S+K+H, Params (M)=5.42023.032.441.23
AnyFlow + GMATraining=C+T+S+K+H, Params (M)=6.02023.032.461.21
GMATraining Data=C + T + S/K (+ H), Condition=Fine-tuned, warm-start=true2021.042.471.39
GMATraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.032.471.39
GMATraining dataset=C+T+S+K+H, warm-start=true2022.042.471.39
GMATraining Data=C+T+S+K+H, warm-start strategy=true2022.032.471.39
GMATraining Data=C+T+S+K+H, warm-start strategy=true2023.032.471.39
GMATraining=C+T+S+K+H, Params (M)=5.92023.032.471.39
AGFlowTraining=C+T+S+K+H, Params (M)=5.62023.032.471.43
GMATraining Data=(C+)T+S+K+H, Warm-start strategy=true2023.112.471.39
AGFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=true2023.112.471.43
GMATraining=C+T+S+K(+H)2024.062.471.39
AGFlowTraining=C+T+S+K(+H)2024.062.471.43
GMA (position-only)Training Data=C + T + S/K (+ H), Condition=Fine-tuned, warm-start=true2021.042.561.48
RAFT + OCTCTraining dataset=C+T+S+K+H, trained on KITTI-2012 and KITTI-2015 datasets=true, warm-start=true2022.042.571.41
MFCFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.581.49
GMA (joint position+content)Training Data=C + T + S/K (+ H), Condition=Fine-tuned, warm-start=true2021.042.631.54
AnyFlow (R.E.)Training=C+T+S+K+H, Params (M)=5.42023.032.631.26
OCAIwarm-start=true2024.032.63—
MatchFlowTraining=C+T+S+K(+H)2024.062.641.33
GMFlowNetTraining Data=C+T+S+K+H2023.032.651.39
GMFlowNetTraining=C+T+S+K+H, Params (M)=9.32023.032.651.39
GMFlowNetTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.651.39
GMFlowNetTraining=C+T+S+K(+H)2024.062.651.39
Separable FlowTraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.032.671.5
S-FlowTraining Data=C+T+S+K+H2022.032.671.5
S-FlowTraining Data=C+T+S+K+H2023.032.671.5
Separable FlowTraining=C+T+S+K+H, Params (M)=6.02023.032.671.5
Separable FlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.671.5
Separable FlowTraining=C+T+S+K(+H)2024.062.671.5
DistractFlowTraining Setting=Semi-Supervised / Additional dataset, Model=RAFT [30], warm-start=true2023.032.71—
DistractFlowwarm-start=true2024.032.71—
DistractFlowTraining Setting=Supervised, Model=RAFT [30], warm-start=true2023.032.77—
FlowSupervisorTraining Setting=Semi-Supervised / Additional dataset, warm-start=true2023.032.79—
FlowSupervisorwarm-start=true2024.032.79—
MFRFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.81.55
DIPTraining=C+T+S+K+H, Params (M)=5.42023.032.831.44
DIPTraining=C+T+S+K(+H)2024.062.831.44
RAFTTraining Data=C + T + S/K (+ H), Condition=Fine-tuned, warm-start=true2021.042.861.61
RAFTTraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.032.861.61
RAFTTraining dataset=C+T+S+K+H, warm-start=true2022.042.861.61
RAFTTraining Data=C+T+S+K+H, warm-start strategy=true2022.032.861.61
RAFTTraining Data=C+T+S+K+H, warm-start strategy=true2023.032.861.61
RAFTTraining=C+T+S+K+H, Params (M)=5.32023.032.861.61
RAFTTraining Setting=Supervised, warm-start=true2023.032.86—
SupervisedTraining Setting=Semi-Supervised / Additional dataset, warm-start=true2023.032.86—
RAFTTraining Data=(C+)T+S+K+H, Warm-start strategy=true2023.112.861.61
RAFT-Supervisedwarm-start=true2024.032.86—
RAFTTraining=C+T+S+K(+H)2024.062.861.61
GMATraining Data=C+T+S+K+H2022.032.881.4
GMATraining Data=C+T+S+K+H, warm-start strategy=false2023.032.881.4
RFPMTraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.032.91.41
GMFlowTraining Data=C+T+S+K+H2023.032.91.74
GMFlowTraining=C+T+S+K+H, Params (M)=4.72023.032.91.74
GMFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.112.91.74
GMFlowTraining=C+T+S+K(+H)2024.062.91.74
RAFT + OCTCTraining dataset=C+T+S+K+H, trained including test images as unlabeled data=true2022.042.951.58
RAFT + OCTCTraining dataset=C+T+S+K+H, trained on KITTI-2012 and KITTI-2015 datasets=true2022.042.981.55
RAFT + OCTCTraining dataset=C+T+S+K+H2022.043.091.82
RAFT-OCTCTraining Setting=Semi-Supervised / Additional dataset, Model=RAFT [30], warm-start=false2023.033.09—
RAFT-OCTCwarm-start=false2024.033.09—
RAFT-ATraining Data=C + T + S/K + H, Evaluation Strategy=warm-start2022.033.142.01
RAFT-ATraining dataset=A+T+S+K+H2022.043.142.01
RAFT-ATraining Setting=Semi-Supervised / Additional dataset, warm-start=false2023.033.14—
RAFT-Awarm-start=false2024.033.14—
RAFTTraining dataset=C+T+S+K+H2022.043.181.94
RAFTTraining Data=C+T+S+K+H2022.033.181.94
RAFTTraining Data=C+T+S+K+H, warm-start strategy=false2023.033.181.94
RAFT-Supervisedwarm-start=false2024.033.18—
FM-RAFTTraining Data=C+T+S+K+H2022.033.61.72
FM-RAFTTraining Data=C+T+S+K+H2023.033.61.72
SCVTraining=C+T+S+K+H, Params (M)=5.32023.033.61.72
StarFlowTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.113.712.72
Flow1DTraining=C+T+S+K+H, Params (M)=5.72023.033.812.24
ScopeFlowTraining dataset=C+T+S+K2022.044.13.59
MaskFlowNetTraining Data=C + T + S/K (+ H), Condition=Fine-tuned2021.044.172.52
MaskFlowNetTraining dataset=C+T+S+K+H2022.044.172.52
MaskFlowNetTraining Data=C+T+S+K+H2022.034.172.52
MaskFlowNetTraining Data=C+T+S+K+H2023.034.172.52
MaskFlowNetTraining Data=(C+)T+S+K+H, Warm-start strategy=false2023.114.172.52