Metric: Accuracy (5-fold) (higher is better)
| # | Model↕ | Accuracy (5-fold)▼ | Augmentations | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | MiVOLO-V2 | 97.39 | Yes | Beyond Specialization: Assessing the Capabilitie... | 2024-03-04 | Code |
| 2 | ViT-hSeq | 96.56 | No | A Hybrid Transformer-Sequencer approach for Age ... | 2024-03-19 | - |
| 3 | MiVOLO-D1 | 96.51 | Yes | MiVOLO: Multi-input Transformer for Age and Gend... | 2023-07-10 | Code |
| 4 | RetinaFace + ArcFace + MLP + Skip connections | 90.66 | Yes | Generalizing MLPs With Dropouts, Batch Normaliza... | 2021-08-18 | Code |
| 5 | CPG (single crop, pytorch) | 89.66 | Yes | Compacting, Picking and Growing for Unforgetting... | 2019-10-15 | Code |
| 6 | PAENet (single crop, tensorflow) | 89.08 | Yes | - | - | Code |
| 7 | Levi_Hassner CNN ( over-sample, caffe) | 86.8 | No | - | - | Code |
| 8 | Levi_Hassner CNN (single crop, caffe) | 85.9 | No | - | - | Code |
| 9 | LMTCNN-2-1 (single crop, tensorflow) | 85.16 | No | Joint Estimation of Age and Gender from Unconstr... | 2018-06-06 | Code |
| 10 | Levi_Hassner CNN (single crop, tensorflow) | 82.52 | No | - | - | Code |