Gaze Estimation on EDIAP (test)
7.1MAE (Img)RAT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RATBackbone=ResNet-18, Training Dataset=G360I Face, Sample size=64002025.02 | 7.1 | — | |
| RATBackbone=ResNet-50, Training Dataset=G360I Face, Sample size=64002025.02 | 7.1 | — | |
| CDGBackbone=ResNet-50, Training Dataset=G360I Face2025.02 | 7.3 | — | |
| Liu et al.Backbone=ResNet-50, Training Dataset=G360I Face2025.02 | 7.5 | — | |
| ST-WSGE (GaT)Backbone=Gaze Transformer, Training Dataset=G360I&V+GF2025.02 | 8.87 | 8.19 | |
| Supervised (GaT)Backbone=Gaze Transformer, Training Dataset=G360I&V2025.02 | 8.88 | 8.28 | |
| Liu et al.Backbone=ResNet-18, Training Dataset=G360I Face2025.02 | 9 | — | |
| PureGazeBackbone=ResNet-18, Training Dataset=G360I Face2025.02 | 9.2 | — |