Research map / Video / Video-language understanding
AMEGO: Active Memory from long EGOcentric videos
Gabriele Goletto, Tushar Nagarajan, Giuseppe Averta, Dima Damen
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 |
|---|---|---|---|
| use AMEGO structured active memory instead of semantic QA baselines | AMEGO · AMB | Total accuracy (%): 23.6 -> 36.3 | Improved |
| use larger visual feature extractors (ViT-L instead of ViT-S/B) (ablation) | AMEGO · AMB | Total accuracy (%): 33.8 -> 36.3 | Improved |
| use a two-stage LLM pipeline (LLoVi) instead of single-stage semantic QA | S-QA baseline · AMB | Total accuracy (%): 22.9 -> 22.6 | Hurt |
| add active-object features to semantic-free QA | SF-QA · AMB | Total accuracy (%): 22.0 -> 21.2 | No clear effect |
| require several detections before starting a tracklet (s_o=20 instead of 1) (ablation) | AMEGO HOI tracking · HOI tracklets standalone | AIoU P: 0.08 -> 0.20 | Improved |
| use hand detections to terminate tracklets (ablation) | AMEGO HOI tracking · HOI tracklets standalone | AIoU P: 0.19 -> 0.20 | Improved |
| add SOT tracker (EgoSTARK) to HOI tracking (ablation) | AMEGO HOI tracking · HOI tracklets standalone | AIoU GT: 0.39 -> 0.41 | Improved |
| add optical flow filter for location segments (ablation) | AMEGO location segments · Location segments standalone | AIoU GT: 0.35 -> 0.50 | Improved |
| add hand detection filter for location segments (ablation) | AMEGO location segments · Location segments standalone | AIoU P: 0.34 -> 0.36 | Improved |
| vary IoU threshold for spatial matching (0.3, 0.5, 0.7) (ablation) | AMEGO HOI tracking · HOI tracklets standalone | AIoU GT: 0.37 / 0.41 / 0.32 | No clear effect |
| vary clustering thresholds theta and tau (ablation) | AMEGO instance assignment · Two annotated videos | ID-switch | No clear effect |
Later papers that compared with this method
1 later accepted paper reports AMEGO as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Building a Mind Palace: Structuring Environment-Grounded Semantic Graphs for Effective Long Video Analysis with LLMs (CVPR 2025 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MMEgo: Towards Building Egocentric Multimodal LLMs for Video QA (ICLR 2025 · similarity 0.90)
- EgoEverything: A Benchmark for Human Behavior–Inspired Long-Context Egocentric Video Understanding in AR Environment (ECCV 2026 · similarity 0.90)
- Omnia de EgoTempo: Benchmarking Temporal Understanding of Multi-Modal LLMs in Egocentric Videos (CVPR 2025 · similarity 0.89)
- EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs (NeurIPS 2025 · similarity 0.88)
- Exocentric-to-Egocentric Video Generation (NeurIPS 2024 · similarity 0.87)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.87)
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.
