Gaze Target Estimation on ChildPlay (test)
95.1AUCGaze-LLE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Gaze-LLEbackbone=ViT-L, without finetuning=true2024.12 | 95.1 | 0.101 | — | |
| Gaze-LLEbackbone=ViT-B, without finetuning=true2024.12 | 94.6 | 0.114 | — | |
| GazeMoE2026.03 | 94.5 | 0.106 | 99.4 | |
| Gaze-LLEBackbone=ViT-L, Evaluation Note=Tested on released Gaze-LLE models on the test set that the authors of ChildPlay dataset suggested2026.03 | 94.2 | 0.111 | 99.3 | |
| Gaze-LLEBackbone=ViT-B, Evaluation Note=Tested on released Gaze-LLE models on the test set that the authors of ChildPlay dataset suggested2026.03 | 93.9 | 0.123 | 99.2 | |
| Tafasca et al.2026.03 | 93.5 | 0.107 | 98.6 | |
| Miao et al.without finetuning=true2024.12 | 93.3 | 0.113 | — | |
| Tafasca et al.without finetuning=true2024.12 | 93.2 | 0.115 | — | |
| Gupta et al.without finetuning=true2024.12 | 92.3 | 0.142 | — | |
| Gupta et al.2026.03 | 91.9 | 0.113 | 98.3 | |
| Chong et al.without finetuning=true2024.12 | 91.2 | 0.121 | — |