Research map / Video / Action recognition
The EPIC-KITCHENS Dataset: Collection, Challenges and Baselines
Dima Damen, Hazel Doughty, Giovanni Maria Farinella, Sanja Fidler, Antonino Furnari, Evangelos Kazakos, Davide Moltisanti, Jonathan Munro, Toby Perrett, Will Price, Michael Wray
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 explicit temporal modelling (TRN, TSM) to TSN-style recognition (ablation) | TSN · EPIC-KITCHENS action recognition (S1) | top-1 verb accuracy: TSN 54.70 -> TRN 61.12, TSM 62.37 | Improved |
| add explicit temporal modelling (TRN, TSM) to TSN-style recognition (ablation) | TSN · EPIC-KITCHENS action recognition (S2) | top-1 verb accuracy: TSN 46.06 -> TRN 51.62, TSM 51.96 | Improved |
| add temporal modelling to recognition for nouns (ablation) | TSN · EPIC-KITCHENS action recognition (S1) | top-1 noun accuracy: TSN 40.11 vs TRN 39.28 vs TSM 41.88 | No clear effect |
| use several snippets (TSN) instead of one snippet (2SCNN) (ablation) | 2SCNN · EPIC-KITCHENS action recognition (S1) | top-1 action accuracy: 13.30 -> 25.43 | Improved |
| fuse RGB, Flow and Audio instead of RGB alone (ablation) | TSN (RGB) · EPIC-KITCHENS action recognition (S1) | top-1 action accuracy: 22.44 -> 26.06 | Improved |
| fuse RGB, Flow and Audio instead of RGB alone (ablation) | TSN (RGB) · EPIC-KITCHENS action recognition (S2) | top-1 action accuracy: 11.30 -> 15.61 | Improved |
| use Flow instead of RGB on unseen kitchens | TSN · EPIC-KITCHENS action recognition (S2) | top-1 noun accuracy: RGB 47.25 vs Flow 42.95 top-5 noun; top-1 verb 36.50 -> 47.49 | Mixed |
| use ResNet-50 backbone for TRN (TRN*) instead of BN-Inception (ablation) | TRN · EPIC-KITCHENS action recognition (S1) | top-1 action accuracy: 27.86 -> 29.41 | Improved |
| use TSM instead of TRN | TRN · EPIC-KITCHENS action recognition (S2) | top-1 action accuracy: TRN 17.34, TSM 17.38 (S2); S1 27.86 -> 29.90 | No clear effect |
| add time-of-day knowledge to majority-class baseline | Largest Class · EPIC-KITCHENS action recognition (S1) | top-1 action accuracy: 02.16 -> 03.57 | Improved |
| use DMR (regress future features) instead of recognition models for anticipation | TSN · EPIC-KITCHENS action anticipation (S1) | top-1 action accuracy: 07.49 -> 01.27 | Hurt |
| add encoder-decoder LSTM (ED) for anticipation | TSN · EPIC-KITCHENS action anticipation (S1) | top-1 action accuracy: 07.49 -> 08.08 (S1); 02.94 -> 02.65 (S2) | Mixed |
Closest papers by meaning
- Scaling Egocentric Vision: The EPIC-KITCHENS Dataset (ECCV 2018 · similarity 0.95)
- HD-EPIC: A Highly-Detailed Egocentric Video Dataset (CVPR 2025 · similarity 0.92)
- Rescaling Egocentric Vision: Collection, Pipeline and Challenges for EPIC-KITCHENS-100 (IJCV 2021 · similarity 0.91)
- CaptainCook4D: A Dataset for Understanding Errors in Procedural Activities (NeurIPS 2024 · similarity 0.85)
- Understanding Multi-Task Activities from Single-Task Videos (CVPR 2025 · similarity 0.85)
- Exploring Domain Incremental Video Highlights Detection with the LiveFood Benchmark (AAAI 2024 · similarity 0.83)
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.
