Research map / Video / Video-language understanding
Egocentric Procedure Parsing
Anubhav Anubhav, Archit Kambhamettu, Vatsal Agarwal, Pulkit Kumar, Abhinav Shrivastava
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 training-free VidParse instead of online ProTAS | online procedure parsing · EgoPER | F1@0.5: 51.31 -> 77.6 | Improved |
| use VidParse instead of online ProTAS | online procedure parsing · GTEA | F1@0.5: 54.87 -> 80.5 | Improved |
| use manipulation-anchored features instead of full-frame DINO CLS (ablation) | VidParse · EgoPER overall | F1: 61.42 -> 77.61 | Improved |
| swap I3D for DINO MAFs in ProTAS (ablation) | ProTAS · EgoPER overall | F1: 51.31 -> 67.26 | Improved |
| increase beam width from 1 to 10 (ablation) | VidParse · EgoPER | F1@0.5: 52.1 -> 77.6 | Improved |
| use centroid micro-prototypes instead of global mean (ablation) | VidParse · EgoPER | F1: 74.5 -> 77.6 | Improved |
| use checkerboard kernel boundary detector instead of z-score (ablation) | VidParse · EgoPER | F1: 69.9 -> 77.6 | Improved |
| vary number of prototypes per action 3 to 11 (ablation) | VidParse · EgoPER | F1: 75.8 -> 77.1 (p=3 vs 11) | No clear effect |
| increase minimum boundary gap d from 20 to 30 (ablation) | VidParse · EgoPER | F1: 77.6 -> 71.3 | Hurt |
| use VidParse instead of ProTAS for inference speed | parsing pipeline · EgoPER | FPS: 4.5 -> 8.6 | Improved |
Closest papers by meaning
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.90)
- Progress-Aware Online Action Segmentation for Egocentric Procedural Task Videos (CVPR 2024 · similarity 0.90)
- Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning (ICLR 2025 · similarity 0.89)
- EgoExoLearn: A Dataset for Bridging Asynchronous Ego- and Exo-centric View of Procedural Activities in Real World (CVPR 2024 · similarity 0.88)
- Learning to Evolve Scenes: Reasoning about Human Activities with Scene Graphs (NeurIPS 2026 · similarity 0.88)
- Do Egocentric Video-Language Models Truly Understand Hand-Object Interactions? (ICLR 2025 · similarity 0.88)
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.
