OpenCodePapers

face-identification-on-megaface

3D Face ReconstructionFacial Recognition and ModellingFace Identification
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
Click a sortable column header to sort.
PaperCodeAccuracyModelNameReleaseDate
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition✓ Link99.27%Cos+UNPG2022-03-22
Partial FC: Training 10 Million Identities on a Single Machine✓ Link99.10%PartialFC + Glint360K + R1002020-10-11
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition✓ Link98.82%Arc+UNPG2022-03-22
Deep Polynomial Neural Networks✓ Link98.78%Prodpoly2020-06-20
GhostFaceNets: Lightweight Face Recognition Model From Cheap Operations✓ Link98.64%GhostFaceNetV2-12023-04-10
ArcFace: Additive Angular Margin Loss for Deep Face Recognition✓ Link98.35%ArcFace + MS1MV2 + R100 + R2018-01-23
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition✓ Link98.03%Mag+UNPG2022-03-22
Support Vector Guided Softmax Loss for Face Recognition✓ Link97.2%SV-AM-Softmax2018-12-29
CosFace: Large Margin Cosine Loss for Deep Face Recognition✓ Link82.72%CosFace2018-01-29
SphereFace: Deep Hypersphere Embedding for Face Recognition✓ Link75.766%SphereFace (3-patch ensemble)2017-04-26
A Light CNN for Deep Face Representation with Noisy Labels✓ Link73.749%Light CNN-292015-11-09
SphereFace: Deep Hypersphere Embedding for Face Recognition✓ Link72.729%SphereFace (single model)2017-04-26
FaceNet: A Unified Embedding for Face Recognition and Clustering✓ Link70.49%FaceNet2015-03-12
Check out how to contribute new results.