Research map / Label-efficient and robust learning / Recognition and long-tail
Ego-PMOVE: Prompt-aware Mixture of View Experts Network for Egocentric Gaze Prediction
Heqian Qiu, Lanxiao Wang, Taijin Zhao, Zhaofeng Shi, Xiang Li, Linfeng Xu, Hongliang Li
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add exocentric videos as plain co-training data (ablation) | CLIP-ViT ego baseline · EgoMe (val) | F1: 49.84 -> 50.06 | No clear effect |
| add MOVE module (ablation) | CLIP-ViT ego baseline with exo data · EgoMe (val) | F1: 50.06 -> 52.08 | Improved |
| add PDM prompt-aware disentangle module (ablation) | Exo + MOVE + VQGD · EgoMe (val) | F1: 52.95 -> 53.44 | Improved |
| replace MOVE with vanilla linear or self-attention fusion (ablation) | Ego-PMOVE · EgoMe (val) | F1: 52.45 -> 50.15 (linear), 50.40 (self-attention) | Hurt |
| add view-discrepancy expert (ablation) | MOVE · EgoMe (val) | F1: 51.49 -> 52.45 | Improved |
| add egocentric-compensated expert (ablation) | MOVE · EgoMe (val) | F1: 52.13 -> 52.45 | Improved |
| add view-shared expert (ablation) | MOVE · EgoMe (val) | F1: 51.68 -> 52.45 | Improved |
| add soft router (ablation) | MOVE · EgoMe (val) | F1: 51.74 -> 52.45 | Improved |
| use multi-scale instead of single-scale gaze decoder (ablation) | gaze decoder · EgoMe (val) | F1: 51.93 -> 52.60 | Improved |
| add view queries to multi-scale decoder (ablation) | gaze decoder · EgoMe (val) | F1: 52.60 -> 52.93 | Improved |
| add either shared loss or discrepancy loss alone (ablation) | KL loss · EgoMe (val) | F1: 52.66 -> 52.94 / 52.97 | No clear effect |
| use full cross-view contrastive loss (shared plus discrepancy) (ablation) | KL loss · EgoMe (val) | F1: 52.66 -> 53.44 | Improved |
Closest papers by meaning
- Gaze-VLM: Bridging Gaze and VLMs through Attention Regularization for Egocentric Understanding (NeurIPS 2025 · similarity 0.90)
- GazeFlow: From Human Gaze Behavior to Generative Egocentric Gaze Prediction (NeurIPS 2026 · similarity 0.89)
- Gaze Beyond the Frame: Forecasting Egocentric 3D Visual Span (NeurIPS 2025 · similarity 0.89)
- In the Eye of MLLM: Benchmarking Egocentric Video Intent Understanding with Gaze-Guided Prompting (NeurIPS 2025 · similarity 0.88)
- Eyes on VLM: Benchmarking Gaze Following and Social Gaze Prediction in Vision Language Models (NeurIPS 2026 · similarity 0.87)
- CLIP-Gaze: Towards General Gaze Estimation via Visual-Linguistic Model (AAAI 2024 · similarity 0.86)
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.
