AI Paper Atlas

Research map / Video / Video-language understanding

Learning Egocentric Cues from Exocentric Video using Privileged Egocentric Supervision

Dominick Reilly, Manish Govind, Le Xue, Srijan Das

ECCV 2026 · Video › Video-language understanding · approach: Egocentric video understanding

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Ego2Exo sequence distillation (ablation)Exo-only VideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 65.7 -> 68.5Improved
add Ego Adaptive Visual Tokens (ablation)Exo-only VideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 65.7 -> 72.2Improved
add both distillation and tokens (ablation)Exo-only VideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 65.7 -> 74.2Improved
add ego videos to training data without distillation (ablation)Exo-only VideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 65.7 -> 66.0No clear effect
use cross-attention to vision encoder features instead of vision-encoder self-attention for tokens (ablation)Ego2ExoVLM · Ego-in-Exo Perceptionavg accuracy (%): 69.9 -> 74.2Improved
train tokens jointly instead of initializing from ego-pretrained cross-attention (ablation)Ego2ExoVLM · Ego-in-Exo Perceptionavg accuracy (%): 73.4 -> 74.2Improved
use language-level sequence distillation instead of visual feature distillation (ablation)Ego2ExoVLM · Ego-in-Exo Perceptionavg accuracy (%): 67.8 -> 74.2Improved
use ego-to-exo language distillation instead of Exo2EgoVLM (ablation)Ego2ExoVLM · Ego-in-Exo Perceptionavg accuracy (%): 70.3 -> 74.2Improved
train on ego-pretrained LLAVIDAL-style token transfer instead (ablation)Ego2ExoVLM · Ego-in-Exo Perceptionavg accuracy (%): 65.4 -> 74.2Improved
train with ego-exo distillation instead of exo-only fine-tuning on same dataVideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 65.7 -> 74.2Improved
add Ego2ExoVLM to VideoLLaMA3VideoLLaMA3 · Ego-in-Exo Perceptionavg accuracy (%): 71.2 -> 74.2Improved
add Ego2ExoVLM to weaker CLIP/Mistral backboneCLIP-Large / Mistral v0.2 · Ego-in-Exo Perceptionavg accuracy (%): 37.0 -> 45.7Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy