List your company
ResearchFormerly WizWand

Face Verification on AgeDB-30 (test)

98.82AccuracyResNet200-TopoFR
92.819294.377195.93597.4929Aug 24, 2019Sep 30, 2020Nov 8, 2021Dec 16, 2022Jan 24, 2024Mar 2, 2025Apr 10, 2026
Updated 6mo ago

Evaluation results

MethodLinks
ResNet200-TopoFRYear='24, Param (M)=118.8, FLOPs (M)=23.5G, Training Dataset=Glint360K, Latency (ms)=12.242026.0498.82
TransFace-LYear='25, Param (M)=271.6, FLOPs (M)=25.4G, Training Dataset=Glint360K, Latency (ms)=OOM2026.0498.66
TransFace-BYear='25, Param (M)=124.5, FLOPs (M)=11.5G, Training Dataset=Glint360K, Latency (ms)=18.202026.0498.62
TransFace-SYear='23, Param (M)=86.7, FLOPs (M)=5.8G, Training Dataset=Glint360K, Latency (ms)=14.312026.0498.5
Mag+UNPGBackbone=ResNet-1002022.0398.38
CurricularFaceBackbone=ResNet101, Training Dataset=MS1MV22021.0598.32
Cos+UNPGBackbone=ResNet-1002022.0398.31
MagFaceBackbone=ResNet-1002022.0398.3
GroupFaceBackbone=ResNet101, Training Dataset=MS1MV22021.0598.28
ResNet50-ArcFaceYear='22, Param (M)=43.6, FLOPs (M)=6.3G, Training Dataset=Glint360K, Latency (ms)=3.762026.0498.28
Arc+UNPGBackbone=ResNet-1002022.0398.25
DCQBackbone=ResNet101, Training Dataset=MS1MV22021.0598.23
ArcFaceBackbone=ResNet-1002022.0398.23
CosFace (ours)Backbone=ResNet101, Training Dataset=MS1MV22021.0598.22
MagFace*Backbone=ResNet-1002022.0398.17
PartialFC-r0.1Backbone=ResNet101, Training Dataset=MS1MV2, Sampling Rate=0.12021.0598.13
CosFaceBackbone=ResNet-1002022.0398.11
FaceLiVTv2-MParam (M)=7.02, FLOPs (M)=258, Training Dataset=Glint360K, Latency (ms)=0.65(18.8× ↓)2026.0498.1
PartialFC-r1.0Backbone=ResNet101, Training Dataset=MS1MV2, Sampling Rate=1.02021.0598.03
FaceLiVTv2-LParam (M)=8.52, FLOPs (M)=309, Training Dataset=Glint360K, Latency (ms)=0.71(17.2× ↓)2026.0498.02
FaceLiVTv2-SParam (M)=4.62, FLOPs (M)=179, Training Dataset=Glint360K, Latency (ms)=0.54(22.7× ↓)2026.0497.82
Seesaw-shuffleFaceNetBatch size=192, Training epoch=162019.0896.85
FaceLiVTv2-XSParam (M)=2.9, FLOPs (M)=90, Training Dataset=Glint360K, Latency (ms)=0.43(28.5× ↓)2026.0496.68
Seesaw-shuffleFaceNet(mobi)Batch size=224, Training epoch=162019.0896.48
Seesaw-shareFaceNetBatch size=160, Training epoch=162019.0896.42
MobileFaceNet (ours)Batch size=256, Training epoch=242019.0896.32
Seesaw-shuffleFaceNetBatch size=160, Training epoch=162019.0896.3
MobileFaceNet#Image=3.8M, Batch size=5122019.0896.07
LMobileNetEBatch size=512, Params=26.7M2019.0896.06
MobileFaceNet (ours)Batch size=192, Training epoch=162019.0896.03
MobileFaceNet (ours)Batch size=256, Training epoch=162019.0896.03
MobileFaceNet (ours)Batch size=160, Training epoch=162019.0895.92
MobileFaceNet#Image=0.5M, Batch size=5122019.0893.05