List your company
ResearchFormerly WizWand

Video Frame Interpolation on UCF101

35.44PSNROurs-Aug (DQBC)
32.496833.260934.02534.7891Oct 20, 2018Jul 21, 2019Apr 21, 2020Jan 21, 2021Oct 23, 2021Jul 25, 2022Apr 26, 2023
Updated 6mo ago

Evaluation results

MethodLinks
Ours-Aug (DQBC)Test augmentation=true2023.0435.4497
Ours-tripletstraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=28.9, Runtime (ms)=6802022.0335.4397.9
NCM (Ours-Large)2022.0735.4397
VFIformer2023.0435.4397
VFIformer2022.0535.4397
RIFE-Large2022.0735.4197
EBME-H*parameters (millions)=3.9, runtime (seconds)=0.082022.0635.4197
SoftSplattraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=7.7, Runtime (ms)=1352022.0335.3997
SoftSplatGFLOPS (share)=95, GFLOPS (unshare)=218, Speed (ms/f)=122, Arbitrary Interpolation=true2022.0435.3997
SoftSplatRuntime (ms) 480p=135*, Parameters (Million)=7.7*2022.0735.3997
SoftSplat2022.0635.3995.2
ABME2022.0235.3896.98
ABMEGFLOPS (share)=N/A, GFLOPS (unshare)=549, Speed (ms/f)=497, Arbitrary Interpolation=true2022.0435.3897
ABME2022.0735.3896.98
ABMEparameters (millions)=18.1, runtime (seconds)=0.282022.0635.3897
ABME2023.0435.3896.98
ABME2022.0535.3896.98
CURE2022.0235.3697.05
NCM (Ours-Base)2022.0735.3696.95
EBME-Hparameters (millions)=3.9, runtime (seconds)=0.042022.0635.3596.9
Ours (DQBC)Test augmentation=false2023.0435.3596.96
EBMEparameters (millions)=3.9, runtime (seconds)=0.022022.0635.396.9
SepConv++training dataset=Vimeo-Triplets-Train, self-ensembling=true2022.0335.29—
RIFE-Ltraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=20.9, Runtime (ms)=722022.0335.2996.9
RIFE2023.0435.2996.9
RIFEmodel_variant=Large2022.0535.2996.93
RIFE2022.0735.2896.9
RIFEGFLOPS (share)=N/A, GFLOPS (unshare)=20, Speed (ms/f)=17, Arbitrary Interpolation=true2022.0435.2596.9
XVFI2022.0735.1896.85
XVFIparameters (millions)=5.5, runtime (seconds)=0.12022.0635.1895.2
M2M-PWCGFLOPS (share)=87, GFLOPS (unshare)=< 1, Speed (ms/f)=32, Arbitrary Interpolation=true2022.0435.1797
BMBC2022.0235.1696.89
BMBCtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=11.0, Runtime (ms)=15802022.0335.1596.9
BMBCGFLOPS (share)=441, GFLOPS (unshare)=376, Speed (ms/f)=1213, Arbitrary Interpolation=true2022.0435.1596.9
BMBC2022.0735.1596.89
BMBCparameters (millions)=11, runtime (seconds)=0.822022.0635.1596.9
BMBC2023.0435.1596.89
BMBC2022.0535.1596.89
EDSCtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=8.9, Runtime (ms)=462022.0335.1396.8
M2M-DISGFLOPS (share)=61, GFLOPS (unshare)=< 1, Speed (ms/f)=28, Arbitrary Interpolation=true2022.0435.1396.8
CyclicGen2023.0435.1196.84
CyclicGen2022.0535.1196.84
XVFI2022.0235.0996.85
DSepConvtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=21.8, Runtime (ms)=2362022.0335.0896.9
FeFlowtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=133.62022.0335.0895.7
MEMC-Net*2018.1035.0196.83
MEMC-Nettraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=70.3, Runtime (ms)=1202022.0335.0196.8
DAINtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=24.0, Runtime (ms)=1302022.033596.8
PWDNtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=7.82022.0335—
DAINGFLOPS (share)=712, GFLOPS (unshare)=1308, Speed (ms/f)=977, Arbitrary Interpolation=true2022.043596.8
DAIN2022.0734.9996.83
DAINparameters (millions)=24, runtime (seconds)=0.152022.0634.9996.8
DAIN2023.0434.9996.83
DAIN2022.0534.9996.83
CAINtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=42.8, Runtime (ms)=382022.0334.9896.9
CAINGFLOPS (share)=N/A, GFLOPS (unshare)=29, Speed (ms/f)=47, Arbitrary Interpolation=false2022.0434.9896.9
ECMparameters (millions)=4.72022.0634.9795.1
MEMC-Net2018.1034.9596.79
RRIN2022.0234.9394.96
CAIN2022.0234.9196.9
CAINparameters (millions)=42.8, runtime (seconds)=0.042022.0634.9196.9
CAIN2023.0434.9196.9
CAIN2022.0534.9196.9
AdaCoF2022.0234.996.8
AdaCoFGFLOPS (share)=N/A, GFLOPS (unshare)=117, Speed (ms/f)=36, Arbitrary Interpolation=false2022.0434.996.8
AdaCoF2022.0734.996.8
AdaCoFparameters (millions)=22.9, runtime (seconds)=0.032022.0634.996.8
AdaCoF2023.0434.996.8
AdaCoF2022.0534.996.8
MEMC-Net_sScale=Small2018.1034.8396.76
SepConvLoss=L12018.1034.7896.69
SepConvtraining dataset=proprietary, # Parameters (Million)=21.6, Runtime (ms)=512022.0334.7896.7
SepConvGFLOPS (share)=N/A, GFLOPS (unshare)=93, Speed (ms/f)=101, Arbitrary Interpolation=false2022.0434.7896.7
SepConv2023.0434.7896.69
SepConv2022.0534.7896.69
SepConvLoss=Lf2018.1034.6996.55
ToFlow+MaskMasking=true2018.1034.5896.67
ToFlow2022.0234.5896.67
TOFlowtraining dataset=Vimeo-Triplets-Train, # Parameters (Million)=1.1, Runtime (ms)=722022.0334.5896.7
ToFlow2023.0434.5896.67
ToFlow2022.0534.5896.67
ToFlow2018.1034.5496.66
DVF2018.1034.1296.31
MIND2018.1033.9396.61
EpicFlow2018.1033.7196.35
SPyNet2018.1033.6796.33
VFIT-B# Parameters (M)=29.02021.1133.4497.1
VFIT-S# Parameters (M)=7.52021.1133.3697.1
FLAVR# Parameters (M)=42.42021.1133.3397.1
FLAVRInputs=RGB, Interpolation factor=2x2020.1233.3197.1
FLAVRInputs=R2022.0133.3197.1
VRTInputs=R2022.0133.397
QVIInputs=RGB+Flow, Interpolation factor=2x2020.1232.8997
QVI# Parameters (M)=29.22021.1132.8997
SoftSplat# Parameters (M)=7.72021.1132.8997
QVIInputs=R+F2022.0132.8997
AdaCoFInputs=RGB, Interpolation factor=2x2020.1232.7196.9
AdaCoF# Parameters (M)=21.82021.1132.7196.9
AdaCoFInputs=R2022.0132.7196.9
BMBCInputs=RGB, Interpolation factor=2x2020.1232.6195.5