| QMagFace: Simple and Accurate Quality-Aware Face Recognition | ✓ Link | 97.72% | 96.48 | 94.7 | | 94.7 | QMagFace | 2021-11-26 |
| Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition | ✓ Link | 97.7% | 96.6 | 95.04 | | | Arc+UNPG | 2022-03-22 |
| Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition | ✓ Link | 97.63% | 96.5 | 95.21 | | | Mag+UNPG | 2022-03-22 |
| Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition | ✓ Link | 97.36% | 96.5 | 94.99 | | | Cos+UNPG | 2022-03-22 |
| FacePoseNet: Making a Case for Landmark-Free Face Alignment | ✓ Link | 96.5% | | | | | FPN | 2017-08-24 |
| GhostVLAD for set-based face recognition | ✓ Link | 96.4% | | | | | SE-GV-3-g2 | 2018-10-23 |
| VGGFace2: A dataset for recognising faces across pose and age | ✓ Link | 95.6% | 90.8 | | | | VGGFace2_ft | 2017-10-23 |
| Cluster and Aggregate: Face Recognition with Large Probe Set | ✓ Link | | 96.91 | 95.53 | 92.29 | | CAFace+AdaFace (WebFace4M) | 2022-10-19 |
| Killing Two Birds with One Stone:Efficient and Robust Training of Face Recognition CNNs by Partial FC | ✓ Link | | | 96.71 | | | PartialFC(WebFace42M) | 2022-03-28 |
| AdaFace: Quality Adaptive Margin for Face Recognition | ✓ Link | | | 96.03 | | | AdaFace (WebFace4M) | 2022-04-03 |
| AdaFace: Quality Adaptive Margin for Face Recognition | ✓ Link | | | 95.84 | | | AdaFace (MS1MV3) | 2022-04-03 |
| AdaFace: Quality Adaptive Margin for Face Recognition | ✓ Link | | | 95.67 | | | AdaFace (MS1MV2) | 2022-04-03 |
| Partial FC: Training 10 Million Identities on a Single Machine | ✓ Link | | | 96.1 | | | Partial FC ResNet100 + CosFace (Glint360K, r=1.0) | 2020-10-11 |
| EdgeFace: Efficient Face Recognition Model for Edge Devices | ✓ Link | | | 93.58 | | | EdgeFace-S (gamma=0.5) | 2023-07-04 |
| LVFace: Large Vision model for Face Recogniton | ✓ Link | | | 96.51 | | | LVFace-L (Glint360K) | 2025-01-23 |
| Vision Transformers for Face Recognition Need More Registers | ✓ Link | | | 96.11 | | | ViT-8R | 2026-06-10 |
| TransFace: Calibrating Transformer Training for Face Recognition from a Data-Centric Perspective | ✓ Link | | | 96.80 | | | TransFace-L (WebFace42M) | 2023-08-20 |
| KeyPoint Relative Position Encoding for Face Recognition | ✓ Link | | | 96.55 | | | AdaFace ViT Base KP-RPE (WebFace12M) | 2024-03-21 |
| FaceLiVT: Face Recognition using Linear Vision Transformer with Structural Reparameterization For Mobile Device | | | | 93.7 | | | FaceLiVT-M-(LA, 40 epochs) | 2025-06-12 |
| ARoFace: Alignment Robustness to Improve Low-Quality Face Recognition | ✓ Link | | | 96.35 | | | AdaFace+ARoFace (WebFace4M) | 2024-07-20 |
| CurricularFace: Adaptive Curriculum Learning Loss for Deep Face Recognition | ✓ Link | | | 94.8 | | | CurricularFace (MS1MV2, ResNet100) | 2020-04-01 |
| MagFace: A Universal Representation for Face Recognition and Quality Assessment | ✓ Link | | | 94.51 | | | MagFace+ | 2021-03-11 |
| ElasticFace: Elastic Margin Loss for Deep Face Recognition | ✓ Link | | | 95.43 | | | ElasticFace-Cos+ (MS1MV2) | 2021-09-20 |
| ArcFace: Additive Angular Margin Loss for Deep Face Recognition | ✓ Link | | | 96.02 | | | IBUG-500K, R100, ArcFace | 2018-01-23 |
| Sparsity-Inducing Divergence Losses for Biometric Verification | | | | 96.44 | | | Q-Margin | 2026-06-30 |
| Learning to Attract and Repel: Dual Quality Margin Learning for Face Recognition (DQM-Face) | | | | 95.38 | | | DQM-Face α=0.4 | 2026-09-08 |
| Noise-Robust Face Recognition via Non-target Similarity Distribution Guided Sample Selection | | | | 95.68 | | | Arc+DDLN (R100) | 2026-09-08 |