Multi-frame Video Interpolation on X 4K (test)
32.92PSNRHiFI
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| HiFITraining dataset=Vimeo + X-TRAIN + Raw videos2024.10 | 32.92 | 0.931 | — | |
| VFIMambaTraining Dataset=Vimeo-90K + X-TRAIN2024.07 | 32.15 | 0.9246 | — | |
| HiFITraining dataset=Vimeo + X-TRAIN2024.10 | 32.03 | 0.918 | — | |
| VFIMamba-STraining Dataset=Vimeo-90K + X-TRAIN2024.07 | 31.73 | 0.9238 | — | |
| EMA-VFITraining dataset=Vimeo (+septuplet for X-TEST)2024.10 | 31.46 | 0.916 | — | |
| SGM-VFITraining Dataset=Vimeo-90K + X-TRAIN2024.07 | 31.35 | 0.9179 | — | |
| BiFormerTraining Dataset=Vimeo-90K + X-TRAIN2024.07 | 31.32 | 0.9215 | — | |
| BiFormerTraining dataset=Vimeo + X-TRAIN2024.10 | 31.32 | 0.921 | — | |
| AMT-GTraining Dataset=Vimeo-90K2024.07 | 31.12 | 0.9157 | — | |
| AMT-LTraining Dataset=Vimeo-90K2024.07 | 30.96 | 0.9147 | — | |
| M2M-VFITraining Dataset=Vimeo-90K2024.07 | 30.89 | 0.9138 | — | |
| M2M-PWCInterpolation Factor=x8, Initial Flow Estimator=PWC-Net2022.04 | 30.81 | 0.912 | — | |
| M2MTraining dataset=Vimeo2024.10 | 30.81 | 0.912 | — | |
| UPR-NetTraining dataset=Vimeo2024.10 | 30.68 | 0.909 | — | |
| IFRNet-LTraining Dataset=Vimeo-90K2024.07 | 30.66 | 0.905 | — | |
| FLDRTraining Dataset=X-TRAIN2024.07 | 30.46 | 0.9041 | — | |
| M2M-DISInterpolation Factor=x8, Initial Flow Estimator=DISFlow2022.04 | 30.18 | 0.909 | — | |
| ABMEInterpolation Factor=x82022.04 | 30.16 | 0.879 | — | |
| RIFETraining Dataset=Vimeo-90K2024.07 | 30.13 | 0.8927 | — | |
| XVFI-NetTesting scale depth (Stst)=5, Number of parameters (#P)=5.5M, Runtime (Rt)=0.075s2021.03 | 30.12 | 0.87 | 2.15 | |
| XVFI+Interpolation Factor=x8, Training Dataset=X-TRAIN2022.04 | 30.12 | 0.87 | — | |
| XVFITraining Dataset=X-TRAIN2024.07 | 30.12 | 0.9045 | — | |
| XVFITraining dataset=Vimeo / X-TRAIN2024.10 | 30.12 | 0.87 | — | |
| CUREtype=continuous multi-frame interpolation2022.02 | 30.05 | 0.8998 | — | |
| EMA-VFI-STraining Dataset=Vimeo-90K2024.07 | 29.91 | 0.8951 | — | |
| XVFI-NetTesting scale depth (Stst)=3, Number of parameters (#P)=5.5M, Runtime (Rt)=0.074s2021.03 | 28.86 | 0.858 | 2.67 | |
| DAINTraining setup=Retrained on X-TRAIN, Number of parameters (#P)=24M, Runtime (Rt)=1.375s2021.03 | 27.52 | 0.821 | 3.47 | |
| FILMTraining dataset=Vimeo2024.10 | 26.98 | 0.839 | — | |
| DAINTraining setup=Original pretrained, Iteration ratio (XN)=x9.3, Number of parameters (#P)=24M, Runtime (Rt)=1.375s2021.03 | 26.78 | 0.807 | 3.83 | |
| DAINInterpolation Factor=x82022.04 | 26.78 | 0.807 | — | |
| AdaCoFTraining setup=Retrained on X-TRAIN, Number of parameters (#P)=21.8M, Runtime (Rt)=0.005s2021.03 | 25.81 | 0.772 | 6.42 | |
| SoftSplatInterpolation Factor=x82022.04 | 25.48 | 0.725 | — | |
| FeFlowTraining setup=Retrained on X-TRAIN, Number of parameters (#P)=102.5M, Runtime (Rt)=1.681s2021.03 | 25.16 | 0.783 | 6.54 | |
| RIFEInterpolation Factor=x82022.04 | 24.67 | 0.797 | — | |
| BMBCtype=continuous multi-frame interpolation2022.02 | 24.28 | 0.7777 | — | |
| XVFItype=continuous multi-frame interpolation2022.02 | 24.12 | 0.7566 | — | |
| FeFlowTraining setup=Original pretrained, Iteration ratio (XN)=x5.3, Number of parameters (#P)=102.5M, Runtime (Rt)=1.681s2021.03 | 24 | 0.756 | 6.59 | |
| SepConvInterpolation Factor=x82022.04 | 23.94 | 0.794 | — | |
| AdaCoFTraining setup=Original pretrained, Iteration ratio (XN)=x5.8, Number of parameters (#P)=21.8M, Runtime (Rt)=0.005s2021.03 | 23.9 | 0.727 | 6.89 | |
| AdaCoFInterpolation Factor=x82022.04 | 23.9 | 0.727 | — | |
| CAINInterpolation Factor=x82022.04 | 22.51 | 0.775 | — | |
| ToFlowtype=continuous multi-frame interpolation2022.02 | 22.33 | 0.7259 | — | |
| SuperSloMotype=continuous multi-frame interpolation2022.02 | 19.58 | 0.7099 | — |