List your company
ResearchFormerly WizWand

Facial Landmark Detection on 300-W (Common)

0.0241NMEADA
-0.0921360.6924571.477052.261643Apr 9, 2019Feb 9, 2020Dec 11, 2020Oct 13, 2021Aug 15, 2022Jun 17, 2023Apr 18, 2024
Updated 8mo ago

Evaluation results

MethodLinks
ADAYear=2020, Backbone=Hourglass2021.050.0241——
DAGYear=2020, Backbone=HRNet-W182021.050.0262——
AWingYear=2019, Backbone=Hourglass2021.050.0272——
BarrelNet (ours)Backbone=ResNet-1012021.050.0273——
BarrelNet (ours)Backbone=ResNet-502021.050.0275——
LUVLiYear=2020, Backbone=DU-Net2021.050.0276——
DCFE2021.020.0276——
PIPNetYear=2020, Backbone=ResNet-1012021.050.0278——
PIPNetYear=2020, Backbone=ResNet-502021.050.028——
Baseline (ours)Backbone=ResNet-1012021.050.0281——
Baseline (ours)Backbone=ResNet-502021.050.0283——
BarrelNet (ours)Backbone=ResNet-182021.050.0283——
HG-HSLEYear=2019, Backbone=Hourglass2021.050.0285——
HRNetYear=2019, Backbone=HRNet-W182021.050.0287——
PIPNetYear=2020, Backbone=ResNet-182021.050.0291——
Baseline (ours)Backbone=ResNet-182021.050.0295——
LABYear=2018, Backbone=ResNet-182021.050.0298——
LAB2021.020.0298——
DAN2021.020.0319——
Ours w SANNormalization Protocol=Inter-ocular Normalization, training=style-augmented2019.080.0321——
AVS w/ SANYear=2019, Backbone=ITN-CPM2021.050.0321——
LaplaceKLYear=20192021.050.0328——
SANNormalization Protocol=Inter-ocular Normalization2019.080.0334——
SANYear=2018, Backbone=ITN-CPM2021.050.0334——
SAN2021.020.0334——
CPM2021.020.0339——
ODNYear=2019, Backbone=ResNet-182021.050.0356——
PCD-CNNNormalization Protocol=Inter-ocular Normalization2019.080.0367——
PCD-CNN2021.020.0367——
DCFENormalization Protocol=Inter-pupil Normalization2019.080.0383——
mnv2_rBackbone=MobileNetV2, Loss function=ASM assisted loss2021.020.0388——
mnv2Backbone=MobileNetV2, Loss function=Standard MSE2021.020.0393——
Ours w Res-18Normalization Protocol=Inter-pupil Normalization, training=style-augmented2019.080.0398——
RARNormalization Protocol=Inter-pupil Normalization2019.080.0412——
DSRN2021.020.0412——
LABNormalization Protocol=Inter-pupil Normalization2019.080.042——
RCN+Year=20182021.050.042——
Ours w LABNormalization Protocol=Inter-pupil Normalization, training=style-augmented2019.080.0423——
TSRNormalization Protocol=Inter-pupil Normalization2019.080.0436——
Res-18Normalization Protocol=Inter-pupil Normalization2019.080.0453——
RCNNormalization Protocol=Inter-pupil Normalization2019.080.0467——
RCN2021.020.0467——
CFSSNormalization Protocol=Inter-pupil Normalization2019.080.0473——
TCDCNNormalization Protocol=Inter-pupil Normalization2019.080.048——
ASMNetBackbone=ASMNet, Loss function=ASM assisted loss2021.020.0482——
SeqMTNormalization Protocol=Inter-pupil Normalization2019.080.0484——
LBFNormalization Protocol=Inter-pupil Normalization2019.080.0495——
RDRNormalization Protocol=Inter-ocular Normalization2019.080.0503——
ESRNormalization Protocol=Inter-pupil Normalization2019.080.0528——
PIFANormalization Protocol=Inter-ocular Normalization2019.080.0543——
CFANNormalization Protocol=Inter-pupil Normalization2019.080.055——
SDMNormalization Protocol=Inter-pupil Normalization2019.080.0557——
ASMNet_nrBackbone=ASMNet, Loss function=Standard MSE2021.020.0586——
3DDFANormalization Protocol=Inter-pupil Normalization2019.080.0615——
ADAYear=2020, Backbone=Hourglass2020.032.41——
STARNormalization=inter-ocular2023.062.52——
STAR loss (smooth-l1 distance)Backbone=Hourglass, Param. (M)=13.372023.062.52——
ADNetNormalization=inter-ocular2023.062.53——
ADNetBackbone=Hourglass, Param. (M)=13.372023.062.53——
DTLDNormalization=inter-ocular2023.062.59——
DTLDBackbone=ResNet-18, Param. (M)=13.32023.062.59——
DAGBackbone=HRNet-W182023.062.62——
HIHNormalization=inter-ocular2023.062.65——
HIHBackbone=Hourglass, Param. (M)=22.682023.062.65——
SLLearning Paradigm=Oracle2024.042.68——
3DDEYear=2019, External area data for training=true2020.032.69——
AWingNormalization=inter-ocular2023.062.72——
AwingBackbone=Hourglass, Param. (M)=25.12023.062.72——
AWingYear=2019, Backbone=Hourglass2020.032.72——
SL+CycleGANLearning Paradigm=DA (300W → CariFace) and GZSL (Unseen ArtiFace)2024.042.74——
SLPTNormalization=inter-ocular2023.062.75——
SLPTBackbone=HRNet-W18C-lite, Param. (M)=9.982023.062.75——
DCFE (w/ 3D)extra info=3D information2019.042.76——
AdalossNormalization=Inter-ocular distance (IOD), Initialization=Ground truth bounding box, Input Resolution=256x2562019.082.76——
DCFE (w/ 3D)extra_info=3D information (3D)2019.082.76——
LUVLiNormalization=inter-ocular2023.062.76——
LUVLiBackbone=DU-Net2023.062.76——
LUVLiYear=2020, Backbone=DU-Net2020.032.76——
DCFEYear=2018, External area data for training=true2020.032.76——
PIPNetNormalization=inter-ocular2023.062.78——
PIPNetBackbone=ResNet-101, Param. (M)=45.72023.062.78——
PIPNetBackbone=ResNet-1012020.032.78——
Conditional Face WarpingLearning Paradigm=DA (300W → CariFace) and GZSL (Unseen ArtiFace)2024.042.79——
PIPNetBackbone=ResNet-502020.032.8——
SL+RevGradLearning Paradigm=DA (300W → CariFace) and GZSL (Unseen ArtiFace)2024.042.84——
HG-HSLEYear=2019, Backbone=Hourglass2020.032.85——
HRNetV2backbone=HRNetV2-W182019.042.87——
HRNetV2-W18backbone=HRNetV2-W182019.082.87——
HRNetBackbone=HRNet-W18, Param. (M)=9.72023.062.87——
HRNetYear=2019, Backbone=HRNetV2-W182020.032.87——
SL+FDALearning Paradigm=DA (300W → CariFace) and GZSL (Unseen ArtiFace)2024.042.89——
SL+ASNLearning Paradigm=DA (300W → ArtiFace) and GZSL (Unseen CariFace)2024.042.89——
DU-NetNormalization=inter-ocular2023.062.9——
Conditional Face WarpingLearning Paradigm=DA (300W → ArtiFace) and GZSL (Unseen CariFace)2024.042.9——
TS³Annotation Ratio=100%, Additional Training Data=Unlabeled AFLW data2019.082.91——
PIPNetBackbone=ResNet-182020.032.91——
SL+ASNLearning Paradigm=DA (300W → CariFace) and GZSL (Unseen ArtiFace)2024.042.92——
DeCaFANormalization=inter-ocular2023.062.93——
DeCaFaBackbone=U-Net, Param. (M)=102023.062.93——
DeCaFAYear=2019, Backbone=Cascaded U-net, External area data for training=true2020.032.93——