Face Alignment On Cofw

Metrics

NME (inter-ocular)

Results

Performance results of various models on this benchmark

Model Name	NME (inter-ocular)	Paper Title	Repository
HRNet	3.45	Deep High-Resolution Representation Learning for Visual Recognition	-
SLPT	3.32	Sparse Local Patch Transformer for Robust Face Alignment and Landmarks Inherent Relation Learning	-
Wing (Feng et al., 2018)	5.07	Wing Loss for Robust Facial Landmark Localisation with Convolutional Neural Networks	-
PIPNet (ResNet-101)	3.08%	Pixel-in-Pixel Net: Towards Efficient Facial Landmark Detection in the Wild	-
DenseU-Net + Dual Transformer	-	Stacked Dense U-Nets with Dual Transformers for Robust Face Alignment	-
DTLD+	3.02%	Towards Accurate Facial Landmark Detection via Cascaded Transformers	-
MobileNetV2+KD-Loss	4.11%	Facial Landmark Points Detection Using Knowledge Distillation-Based Neural Networks	-
CHR2C (Inter-pupils Norm)	-	Cascade of Encoder-Decoder CNNs with Learned Coordinates Regressor for Robust Facial Landmarks Detection
LAB (w/ B)	3.92%	Look at Boundary: A Boundary-Aware Face Alignment Algorithm	-
LAB	5.58%	Look at Boundary: A Boundary-Aware Face Alignment Algorithm	-
Ours (VGG-F)	3.32	Pre-training strategies and datasets for facial representation learning	-
ATF	3.32%	ATF: Towards Robust Face Alignment via Leveraging Similarity and Diversity across Different Datasets	-
MNN (Inter-pupil Norm)	-	Multi-task head pose estimation in-the-wild	-
PropNet	3.71%	PropagationNet: Propagate Points to Curve to Learn Structure Information	-
STAR	3.21%	STAR Loss: Reducing Semantic Ambiguity in Facial Landmark Detection	-
BarrelNet (ResNet-101)	3.1%	When Liebig's Barrel Meets Facial Landmark Detection: A Practical Model	-
FiFA	2.96	Fiducial Focus Augmentation for Facial Landmark Detection	-
EF-3ACR	3.47%	ACR Loss: Adaptive Coordinate-based Regression Loss for Face Alignment	-
SCC	3.63%	Fast and Accurate: Structure Coherence Component for Face Alignment	-
DCFE	-	A Deeply-initialized Coarse-to-fine Ensemble of Regression Trees for Face Alignment	-

0 of 28 row(s) selected.