SOTAVerified

Facial Landmark Detection

Facial Landmark Detection is a computer vision task that involves detecting and localizing specific points or landmarks on a face, such as the eyes, nose, mouth, and chin. The goal is to accurately identify these landmarks in images or videos of faces in real-time and use them for various applications, such as face recognition, facial expression analysis, and head pose estimation.

( Image credit: Style Aggregated Network for Facial Landmark Detection )

Papers

Showing 1–10 of 139 papers

TitleStatusHype
FaceXFormer: A Unified Transformer for Facial AnalysisCode3
STAR Loss: Reducing Semantic Ambiguity in Facial Landmark DetectionCode2
Shape Preserving Facial Landmarks with Graph Attention NetworksCode2
KeyPosS: Plug-and-Play Facial Landmark Detection through GPS-Inspired True-Range MultilaterationCode1
HPRNet: Hierarchical Point Regression for Whole-Body Human Pose EstimationCode1
img2pose: Face Alignment and Detection via 6DoF, Face Pose EstimationCode1
Deep Structured Prediction for Facial Landmark DetectionCode1
A Detailed Look At CNN-based Approaches In Facial Landmark DetectionCode1
AnchorFace: An Anchor-based Facial Landmark Detector Across Large PosesCode1
ArtFacePoints: High-resolution Facial Landmark Detection in Paintings and PrintsCode1
Show:102550
← PrevPage 1 of 14Next →

Benchmark Results

#ModelMetricClaimedVerifiedStatus
13DDFANME7.01—Unverified
2Pose-InvariantNME6.3—Unverified
3CFSSNME5.76—Unverified
4SAN GTNME3.98—Unverified
5TS3NME3.49—Unverified
6AdalossNME3.31—Unverified
7CHR2C (Inter-ocular Norm)NME3.3—Unverified
8CNN-CRF (Inter-ocular Norm)NME3.3—Unverified
9DCFE (Inter-ocular Norm)NME3.24—Unverified
103DDE (Inter-ocular Norm)NME3.13—Unverified
#ModelMetricClaimedVerifiedStatus
1HPRNet (Hourglass-104)keypoint AP75.4—Unverified
2HPRNet (DLA)keypoint AP74.6—Unverified
3HPRNet (Hourglass-104)keypoint AP59.4—Unverified
4HPRNet (DLA)keypoint AP55.2—Unverified
5HPRNet (Hourglass-104)keypoint AP50.4—Unverified
6HPRNet (DLA)keypoint AP47—Unverified
#ModelMetricClaimedVerifiedStatus
1DCFE (Box height Norm, 19 landmarks - no earlobs)Mean NME 2.17—Unverified
23DDE (Box height Norm, 19 landmarks - no earlobs)Mean NME2.01—Unverified
3SANMean NME 1.91—Unverified
4AnchorFaceMean NME 1.56—Unverified
5FiFAMean NME 0.92—Unverified
#ModelMetricClaimedVerifiedStatus
1SANMean NME 1.85—Unverified
2AnchorFaceMean NME1.38—Unverified
3FiFAMean NME0.8—Unverified
#ModelMetricClaimedVerifiedStatus
1ELD (MobileNetV2)NME3.09—Unverified
2ELD (EfficientNetV2B0)NME2.98—Unverified
3ELD (EfficientNetV2S)NME2.83—Unverified
#ModelMetricClaimedVerifiedStatus
1ELD (EfficientNetV2B1)NME4.65—Unverified
2FiFANME (inter-ocular)3.89—Unverified
3D-ViTNME3.75—Unverified
#ModelMetricClaimedVerifiedStatus
1CPM+SBR+PAMAUC0.08 private59.39—Unverified
2CPM+SBRAUC0.08 private58.22—Unverified
#ModelMetricClaimedVerifiedStatus
1AnchorFaceMean NME 3.72—Unverified
2TS3Mean NME 3.49—Unverified
#ModelMetricClaimedVerifiedStatus
1D-ViTNME (inter-pupil)4.13—Unverified
2FiFANME (inter-ocular)2.96—Unverified
#ModelMetricClaimedVerifiedStatus
1JVCRGTE7.28—Unverified