SOTA Facial Landmark Detection benchmarks and papers with code

Benchmarks

Dataset Name	SOTA Method	Metric
300-W (Common)		NME0.0241	180	4d ago
300-W (Fullset)	SHT-M-i	Mean Error (%)2.78	174	4d ago
300W challenging	CCNN	NME3.99	159	4d ago
WFLW (test)	SH-FAN	Mean Error (ME) - All3.72	122	4d ago
AFLW Full	AnchorFace	NME0.0156	101	4d ago
COFW (test)	HIH	NME0.0321	93	4d ago
WFLW (Full)	Ours	NME (%)3.72	65	4d ago
300W	STAR loss (smooth-l1 distance)	NME2.87	52	4d ago
WFLW Pose	ADNet	Mean Error (%)6.96	50	4d ago
WFLW Blur	SPIGA	Mean Error (%)4.65	49	4d ago
WFLW Make-up	SPIGA	Mean Error3.81	49	4d ago
300-W Challenging Subset	Structure Coherence Component	Mean Error4.93	49	4d ago
300-W public Challenging inter-pupil normalization (test)	HGs + SA + Norm	NME6.38	46	4d ago
WFLW Illumination	SPIGA	Failure Rate @0.11.58	45	4d ago
WFLW Expression	HIHB	Failure Rate @0.11.27	45	4d ago
AFLW Front	LAB	NME1.14	38	4d ago
300-W Common Subset	SHT-M-i	Mean Error (%)2.46	37	4d ago
300W official (test)	LAB	AUC (0.08)58.85	32	4d ago
MAFL (test)	Deep Latent Particles	Normalised MSE (%)2.42	30	4d ago
300-W public Full inter-pupil normalization (test)	CNN-6/7	NME4.04	29	4d ago
300-W Public Common subset inter-pupil normalization (test)	CNN-6/7	NME3.27	28	4d ago
300-W public (test)	GV(HRNet)	NME2.62	28	4d ago
AFLW Full (test)	LUVLi	Average Error1.39	26	4d ago
COFW (Full)	PIPNet	NME3.08	22	4d ago
300W Fullset 2013 (test)	GV(HRNet)	NME2.99	21	4d ago

Showing 25 of 67 rows