Face Recognition
Facial Recognition is the task of making a positive identification of a face in a photo or video image against a pre-existing database of faces. It begins with detection - distinguishing human faces from other objects in the image - and then works on identification of those detected faces.
The state of the art tables for this task are contained mainly in the consistent parts of the task : the face verification and face identification tasks.
( Image credit: Face Verification )
Papers
Showing 101–125 of 2329 papers
All datasetsLFWCFP-FPCASIA-WebFace+masksCelebA+masksMLFWAgeDB-30Color FERETIJB-BCALFWmebeblurfMFW+ (M-M)MFW+ (U-M)
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | PIC - ArcFace | FNMR [%] @ 10-3 FMR | 4.38 | — | Unverified |
| 2 | GhostFaceNetV2-1 (MS1MV3) | Accuracy | 1 | — | Unverified |
| 3 | SymFace + AdaFace + ResNet100 +WebFace (MS1MV2) | Accuracy | 1 | — | Unverified |
| 4 | Prodpoly | Accuracy | 1 | — | Unverified |
| 5 | DiscFace | Accuracy | 1 | — | Unverified |
| 6 | ArcFace + MS1MV2 + R100 | Accuracy | 1 | — | Unverified |
| 7 | DCQ | Accuracy | 1 | — | Unverified |
| 8 | AdaFace + WebFace4M + R100 | Accuracy | 1 | — | Unverified |
| 9 | EdgeFace - S (g=0.5) | Accuracy | 1 | — | Unverified |
| 10 | EdgeFace - XS (g=0.6) | Accuracy | 1 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GhostFaceNetV2-1 | Accuracy | 0.99 | — | Unverified |
| 2 | SymFace + AdaFace + ResNet100 +WebFace | Accuracy | 0.99 | — | Unverified |
| 3 | ElasticFace-Arc | Accuracy | 0.99 | — | Unverified |
| 4 | DiscFace | Accuracy | 0.99 | — | Unverified |
| 5 | CircleLoss(ours) | Accuracy | 0.96 | — | Unverified |
| 6 | EdgeFace - S (g=0.5) | Accuracy | 0.96 | — | Unverified |
| 7 | DCQ | Accuracy | 0.93 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Fine-tuned ArcFace | Accuracy | 91.47 | — | Unverified |
| 2 | Fine-tuned FaceNet | Accuracy | 88.06 | — | Unverified |
| 3 | ArcFace | Accuracy | 87.95 | — | Unverified |
| 4 | Fine-tuned VGG-Face | Accuracy | 86.85 | — | Unverified |
| 5 | FaceNet | Accuracy | 84.21 | — | Unverified |
| 6 | VGG-Face | Accuracy | 79.65 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Fine-tuned ArcFace | Accuracy | 95.43 | — | Unverified |
| 2 | Fine-tuned FaceNet | Accuracy | 93.58 | — | Unverified |
| 3 | ArcFace | Accuracy | 91.78 | — | Unverified |
| 4 | Fine-tuned VGG-Face | Accuracy | 91.51 | — | Unverified |
| 5 | FaceNet | Accuracy | 90.96 | — | Unverified |
| 6 | VGG-Face | Accuracy | 84.56 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | MS1MV2, R100, SFace | Accuracy | 91.57 | — | Unverified |
| 2 | MS1MV2, R100, Arcface | Accuracy | 90.57 | — | Unverified |
| 3 | MS1MV2, R100, Curricularface | Accuracy | 90.43 | — | Unverified |
| 4 | VGGFace2, R50, ArcFace | Accuracy | 85.95 | — | Unverified |
| 5 | CASIA-WebFace, R50, CosFace | Accuracy | 82.52 | — | Unverified |
| 6 | Private-Asia, R50, ArcFace | Accuracy | 77.2 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Prodpoly | Accuracy | 0.98 | — | Unverified |
| 2 | ElasticFace-Cos | Accuracy | 0.98 | — | Unverified |
| 3 | Transformer loss+ArcFace ResNet100 | Accuracy | 0.98 | — | Unverified |
| 4 | DCQ | Accuracy | 0.98 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | VGG based | 5-class test accuracy | 99.2 | — | Unverified |
| 2 | PIC - ArcFace | FNMR [%] @ 10-3 FMR | 4.22 | — | Unverified |
| 3 | PIC - MagFace | FNMR [%] @ 10-3 FMR | 3.92 | — | Unverified |
| 4 | PIC - QMagFace | FNMR [%] @ 10-3 FMR | 3.24 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | ElasticFace-Cos | TAR @ FAR=0.0001 | 0.95 | — | Unverified |
| 2 | ArcFace+CSFM | Rank-1 | 0.95 | — | Unverified |
| 3 | ArcFace + MS1MV2 + R100 | Rank-1 | 0.95 | — | Unverified |
| 4 | AdaFace + MS1MV3 + R100 | TAR @ FAR=0.0001 | 0.94 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Prodpoly | Accuracy | 0.96 | — | Unverified |
| 2 | ElasticFace-Arc | Accuracy | 0.96 | — | Unverified |
| 3 | GhostFaceNetV2-1 | Accuracy | 0.96 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | PIC - ArcFace | FNMR [%] @ 10-3 FMR | 10.1 | — | Unverified |
| 2 | PIC - QMagFace | FNMR [%] @ 10-3 FMR | 9.78 | — | Unverified |
| 3 | PIC - MagFace | FNMR [%] @ 10-3 FMR | 9.75 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | ArcFace + PPL | TAR@FAR=0.0001 | 78.8 | — | Unverified |
| 2 | MaskInv-HG | TAR@FAR=0.0001 | 78.36 | — | Unverified |
| 3 | FocusFace | TAR@FAR=0.0001 | 77.77 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | AdaFace+PPL | TAR@FAR=0.0001 | 84.27 | — | Unverified |
| 2 | MaskInv-HG | TAR@FAR=0.0001 | 83.82 | — | Unverified |
| 3 | FocusFace | TAR@FAR=0.0001 | 83.4 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | PIC - QMagFace | FNMR [%] @ 10-3 FMR | 0.96 | — | Unverified |
| 2 | PIC - MagFace | FNMR [%] @ 10-3 FMR | 0.96 | — | Unverified |
| 3 | PIC - ArcFace | FNMR [%] @ 10-3 FMR | 0.05 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | FaceNet+Adaptive Threshold | Average Accuracy (10 times) | 84.3 | — | Unverified |
| 2 | FaceNet+Fixed Threshold (0.2487) | Average Accuracy (10 times) | 80.6 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Model with Up Convolution + DoG Filter (Aligned) | Rank-1 | 85 | — | Unverified |
| 2 | DPM | Rank-1 | 71 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | FaceNet+Adaptive Threshold | Average Accuracy (10 times) | 83.79 | — | Unverified |
| 2 | FaceNet+Fixed Threshold (0.3968) | Average Accuracy (10 times) | 80.72 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GhostFaceNetV2-1 | Accuracy | 0.95 | — | Unverified |
| 2 | ElasticFace-Arc | Accuracy | 0.93 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | FaceNet+Adaptive Threshold | Average Accuracy (10 times) | 76.46 | — | Unverified |
| 2 | FaceNet+Fixed Threshold (0.3779) | Average Accuracy (10 times) | 53.97 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Wang et al. [5] | Rank-1 | 94.5 | — | Unverified |
| 2 | Multi-task | Rank-1 | 84.32 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Model with Up Convolution + DoG Filter | Rank-1 | 87.2 | — | Unverified |
| 2 | DPM | Rank-1 | 83.73 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | MCN | TAR @ FAR=0.01 | 0.54 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | GhostFaceNetV2-1 | Accuracy | 99.91 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Model with Up Convolution + DoG Filter | Rank-1 | 88.33 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | Partial FC | MFR-ALL | 97.85 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | FaceTransformer+OctupletLoss | Accuracy | 0.95 | — | Unverified |