List your company
ResearchFormerly WizWand

Image Deepfake Detection on FFIW

0.965AUCBlenD
0.636360.721680.8070.89232Aug 30, 2024Dec 11, 2024Mar 24, 2025Jul 5, 2025Oct 16, 2025Jan 27, 2026May 11, 2026
Updated 5mo ago

Evaluation results

MethodLinks
BlenDTraining Data=25k ScaleDF (real) + SBI fakes, Evaluation Protocol=Video-level2026.050.965
GenD (DINO)Input=Frame, Backbone=DINOv3 ViT-L/16, Training Dataset=FF++2025.080.945
GenD (PE)Input=Frame, Backbone=PEcoreL, Training Dataset=FF++2025.080.933
GenD-PETraining Data=FF++, Evaluation Protocol=Video-level2026.050.933
HAADTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.932
GenD (CLIP)Input=Frame, Backbone=CLIP ViT-L/14, Training Dataset=FF++2025.080.928
EffortTraining Data=FF++, Evaluation Protocol=Video-level2026.050.925
OursTraining Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.921
P&PInput=Video, Backbone=CLIP ViT-L/14, Training Dataset=FF++2025.080.921
EffortInput=Frame, Backbone=CLIP ViT-L/14, Training Dataset=FF++2025.080.921
X^2-DFDTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation, Re-implemented=true2026.050.915
ForAdaTraining Data=FF++, Evaluation Protocol=Video-level2026.050.906
FS-VFMTraining Data=FF++, Evaluation Protocol=Video-level2026.050.906
EffortTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation, Re-implemented=true2026.050.901
VLFFDTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation, Re-implemented=true2026.050.898
M2F2-Det (w/ SBI)Training set (Real)=true, Training set (Fake)=false2025.030.887
DAIDTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation, Re-implemented=true2026.050.883
SBITraining Data=FF++ with pseudo-fakes, Evaluation Protocol=Video-level2026.050.881
LSDAVenues=CVPR 2024, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.879
FIA-USATraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation, Re-implemented=true2026.050.878
SBIVenues=CVPR 2022, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.866
SBITraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.866
DCLVenues=AAAI 2022, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.863
FreqBlenderVenue=NeurIPS24, Training set (Real)=true, Training set (Fake)=false2025.030.8614
SBIVenue=CVPR22, Training set (Real)=true, Training set (Fake)=false2025.030.8483
SBIInput=Frame, Backbone=EFNB4, Training Dataset=FF++2025.080.848
CFMVenues=TIFS 2023, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.831
FSBITraining Data=FF++ with pseudo-fakes, Evaluation Protocol=Video-level2026.050.824
AUNetInput=Video, Backbone=Xception+ART, Training Dataset=FF++2025.080.815
AUNetVenue=CVPR23, Training set (Real)=true, Training set (Fake)=false2025.030.8145
SPSLVenues=CVPR 2021, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.794
SRMVenues=CVPR 2021, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.794
SPSLTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.794
SRMTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.794
FSFMTraining Data=FF++, Evaluation Protocol=Video-level2026.050.782
IIDTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.762
FTCNInput=Video, Backbone=3D ResNet-50, Training Dataset=FF++2025.080.745
LSDATraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.724
LSDAInput=Frame, Backbone=EFNB4, Training Dataset=FF++2025.080.724
RECCEVenues=CVPR 2022, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.711
RECCETraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.711
COREVenues=CVPRW 2022, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.71
CORETraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.71
UCFVenues=ICCV 2023, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.697
UCFTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.697
F3NetVenues=ECCV 2020, Training Dataset=FF++ c23, Evaluation Protocol=Protocol-12024.080.649
F3NetTraining Dataset=FF++, Evaluation Protocol=Cross-dataset Evaluation2026.050.649