Action Recognition on EGTEA Gaze+ (S1/S3/Avg Top-1)
81.8Average Top-1 AccuracyLaViLa
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LaViLaExternal Pre-training=Large-scale / LLM / VLM2026.07 | 81.8 | — | — | |
| EgoVideo-GExternal Pre-training=Large-scale / LLM / VLM2026.07 | 80 | — | — | |
| LaViLa-LExternal Pre-training=Large-scale / LLM / VLM2026.07 | 76 | — | — | |
| Ego-VPAExternal Pre-training=Large-scale / LLM / VLM2026.07 | 73.4 | — | — | |
| TSF-BExternal Pre-training=Large-scale / LLM / VLM2026.07 | 65.6 | — | — | |
| GC-TSM2026.07 | 65.1 | — | 62.6 | |
| SAGE2026.07 | 65 | 61.8 | 63.1 | |
| SAP2026.07 | 62.7 | — | 62 | |
| R34-2StreamBackbone=ResNet-342026.07 | 60.8 | — | 58.6 | |
| I3D-2StreamBackbone=I3D2026.07 | 54.2 | — | 53.6 | |
| EGTEA2026.07 | — | 57.2 | — | |
| MCN2026.07 | — | 55.7 | — |