Metric: Accuracy (higher is better)
| # | Model↕ | Accuracy▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | EfficientFER | 82.47 | Yes | - | - | Code |
| 2 | FERNeXt-SDAFE | 81.33 | Yes | - | - | - |
| 3 | ResEmoteNet | 79.79 | Yes | ResEmoteNet: Bridging Accuracy and Loss Reductio... | 2024-09-01 | Code |
| 4 | Ensemble ResMaskingNet with 6 other CNNs | 76.82 | Yes | - | - | Code |
| 5 | Mini-ResEmoteNet (A) | 76.33 | Yes | Mini-ResEmoteNet: Leveraging Knowledge Distillat... | 2025-01-30 | - |
| 6 | EmoNeXt | 76.12 | No | EmoNeXt: an Adapted ConvNeXt for Facial Emotion ... | 2025-01-14 | Code |
| 7 | Segmentation VGG-19 | 75.97 | Yes | - | - | Code |
| 8 | Local Learning Deep+BOW | 75.42 | Yes | Local Learning with Deep and Handcrafted Feature... | 2018-04-29 | - |
| 9 | LHC-Net | 74.42 | Yes | Local Multi-Head Channel Self-Attention for Faci... | 2021-11-14 | Code |
| 10 | Residual Masking Network | 74.14 | Yes | - | - | Code |
| 11 | ResNet18 With Tricks | 73.7 | No | - | - | Code |
| 12 | VGGNet | 73.28 | No | Facial Emotion Recognition: State of the Art Per... | 2021-05-08 | Code |
| 13 | CNN Hyperparameter Optimisation | 72.16 | No | - | - | Code |
| 14 | Ad-Corre | 72.03 | No | - | - | Code |
| 15 | Mini-ResEmoteNet (B) | 70.2 | Yes | Mini-ResEmoteNet: Leveraging Knowledge Distillat... | 2025-01-30 | - |
| 16 | DeepEmotion | 70.02 | No | Deep-Emotion: Facial Expression Recognition Usin... | 2019-02-04 | Code |
| 17 | Local Learning BOW | 67.48 | No | Challenges in Representation Learning: A report ... | 2013-07-01 | Code |