Egocentric visual attention prediction on Ego4D (test) benchmark leaderboard · SOTA2 Research

Egocentric visual attention prediction on Ego4D (test)

0.401F1 ScoreLanguage-guided scene context-aware learning framework

Evaluation Results

Method
Language-guided scene context-aware learning framework2026.01	0.401	0.541	0.319
CSTS2026.01	0.397	0.533	0.316
GLC2026.01	0.378	0.529	0.294
DFG+2026.01	0.373	0.523	0.29
DFG2026.01	0.372	0.532	0.286
MViT2026.01	0.372	0.541	0.283
AttnTrans2026.01	0.37	0.55	0.279
I3D-R502026.01	0.369	0.521	0.286
GazeMLE2026.01	0.363	0.525	0.278