List your company
ResearchFormerly WizWand

Landmark Localization on AFLW (test)

1.59NME (%)RCN+ (L+ELT+A)
1.18963.89236.5959.2977Sep 5, 2017Oct 23, 2017Dec 10, 2017Jan 27, 2018Mar 16, 2018May 3, 2018Jun 20, 2018
Updated 8mo ago

Evaluation results

MethodLinks
RCN+ (L+ELT+A)Labeled Images=100%2017.091.59
RCN+Loss components=L + ELT + A, Labeled Images percentage=100%2017.091.59
RCN+ (L+ELT)Labeled Images=100%2017.091.6
RCN+Loss components=L + ELT, Labeled Images percentage=100%2017.091.6
RCN+ (L)Labeled Images=100%2017.091.61
RCN+Loss components=Landmark (L), Labeled Images percentage=100%2017.091.61
RCN+ (L+ELT+A)Labeled Images=5%2017.092.03
RCN+Loss components=L + ELT + A, Labeled Images percentage=5%2017.092.03
RCN+ (L+A)Labeled Images=5%2017.092.08
RCN+Loss components=L + ELT, Labeled Images percentage=5%2017.092.08
RCN+Loss components=L + A, Labeled Images percentage=5%2017.092.08
Lv et. alLabeled Images=100%2017.092.17
RCN+ (L)Labeled Images=5%2017.092.17
Lv et. alLabeled Images percentage=100%2017.092.17
RCN+Loss components=Landmark (L), Labeled Images percentage=5%2017.092.17
RCN+ (L+ELT)Labeled Images=5%2017.092.22
RCN+ (L+ELT+A)Labeled Images=1%2017.092.46
RCN+Loss components=L + ELT + A, Labeled Images percentage=1%2017.092.46
RCN+ (L+A)Labeled Images=1%2017.092.52
RCN+Loss components=L + A, Labeled Images percentage=1%2017.092.52
CCLLabeled Images=100%2017.092.72
CCLLabeled Images percentage=100%2017.092.72
RCN+ (L)Labeled Images=1%2017.092.88
RCN+Loss components=Landmark (L), Labeled Images percentage=1%2017.092.88
RCPRLabeled Images=100%2017.093.73
RCPRLabeled Images percentage=100%2017.093.73
CFSSLabeled Images=100%2017.093.92
CFSSLabeled Images percentage=100%2017.093.92
SDMLabeled Images=100%2017.094.05
SDMLabeled Images percentage=100%2017.094.05
LBFLabeled Images=100%2017.094.25
LBFLabeled Images percentage=100%2017.094.25
ERTLabeled Images=100%2017.094.35
ERTLabeled Images percentage=100%2017.094.35
CDMLabeled Images=100%2017.095.43
CDMLabeled Images percentage=100%2017.095.43
Proposed MethodTraining Set=CelebA, Loss Network=self-supervised, K=302018.066.31
Proposed MethodTraining Set=CelebA, Loss Network=self-supervised, K=502018.066.33
Proposed MethodTraining Set=CelebA, Loss Network=supervised, K=502018.066.35
Proposed MethodTraining Set=CelebA, Loss Network=supervised, K=302018.066.39
ZhangSupervision=Unsupervised, K=30, Equivariance=true2018.066.58
Proposed MethodTraining Set=VoxCeleb, Loss Network=self-supervised, K=302018.066.75
Proposed MethodTraining Set=CelebA, Loss Network=self-supervised, K=102018.066.86
MTCNNSupervision=Supervised2018.066.9
Proposed MethodTraining Set=CelebA, Loss Network=supervised, K=102018.066.99
ZhangSupervision=Unsupervised, K=102018.067.01
Proposed MethodTraining Set=VoxCeleb, Loss Network=supervised, K=302018.067.1
RARSupervision=Supervised2018.067.23
TCDCNSupervision=Supervised2018.067.65
Thewlis (frames)Supervision=Unsupervised2018.068.8
Cascaded CNNSupervision=Supervised2018.068.97
ThewlisSupervision=Unsupervised, K=502018.0610.53
CFANSupervision=Supervised2018.0610.94
RCPRSupervision=Supervised2018.0611.6