Gaze target estimation on ChildPlay
0.084L2 Distancemulti-scale object-aware framework
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| multi-scale object-aware frameworkParams=7.1M2026.06 | 0.084 | 0.987 | 0.99 | — | |
| Gaze-LLE (ViT-L)Backbone=ViT-L, Learnable Params=2.9M, Input=I2024.12 | 0.101 | 0.951 | 0.994 | 0.662 | |
| Gaze-LLEParams=2.8M2026.06 | 0.101 | 0.951 | 0.994 | — | |
| Tafasca et al. [61]Learnable Params=105M, Input=I2024.12 | 0.106 | — | 0.99 | 0.6 | |
| Gaze-LLE (ViT-B)Backbone=ViT-B, Learnable Params=2.8M, Input=I2024.12 | 0.106 | 0.949 | 0.994 | 0.715 | |
| SharinganParams=>135M*2026.06 | 0.106 | — | 0.99 | — | |
| Tafasca et al. [60]Learnable Params=>25M*, Input=I+D2024.12 | 0.107 | 0.935 | 0.986 | 0.663 | |
| Tafasca et al.2026.06 | 0.107 | 0.935 | 0.986 | — | |
| Gupta et al.Learnable Params=35M, Input=I+D+P2024.12 | 0.113 | 0.919 | 0.983 | 0.694 | |
| Gupta et al.2026.06 | 0.113 | 0.919 | 0.983 | — | |
| MTGS2026.06 | 0.113 | — | 0.993 | — |