Research map / Video / Action recognition
E3V-K5: An Authentic Benchmark for Redefining Video-Based Energy Expenditure Estimation
Shengxuming Zhang, Lei Jin, Yifan Wang, Xinyu Wang, Xu Wen, Zunlei Feng, Mingli Song
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 category-related joint-specific attention from action recognition branch (ablation) | E3SFormer · E3V-K5 test (w/o MM) | MAE: 2.1071 -> 2.0304 | Improved |
| add joint-specific attention (ablation) | E3SFormer · E3V-K5 test (w/ MM) | MAE: 1.5105 -> 1.2490 | Improved |
| add heart rate and physical attributes as input (ablation) | E3SFormer · E3V-K5 test | MAE: 2.0304 -> 1.2490 | Improved |
| add video skeleton features to heart rate and attributes (ablation) | MLP on HR+Attr · E3V-K5 test | MAE: 1.5276 -> 1.2490 | Improved |
| use learned MLP on heart rate and attributes instead of ACSM formula (ablation) | EE regression from HR and attributes · E3V-K5 test | MAE: 3.5047 -> 1.5276 | Improved |
| use heart rate only or attributes only instead of both (ablation) | MLP regression · E3V-K5 test | MAE: 1.5276 -> 2.7904 (HR only), 3.0140 (Attr only) | Hurt |
| use RGB video frames instead of skeleton rendering as input | Conv3D · E3V-K5 test (w/o MM) | MAE: 2.0670 -> 2.5408 | Hurt |
| E3SFormer vs PoseConv3D on R2 (skeleton only) | E3SFormer · E3V-K5 test (w/o MM) | R2: 0.5175 -> 0.5118 | No clear effect |
Closest papers by meaning
- Jointly Learning Energy Expenditures and Activities Using Egocentric Multimodal Signals (CVPR 2017 · similarity 0.83)
- M3GYM: A Large-Scale Multimodal Multi-view Multi-person Pose Dataset for Fitness Activity Understanding in Real-world Settings (CVPR 2025 · similarity 0.83)
- AIFit: Automatic 3D Human-Interpretable Feedback Models for Fitness Training (CVPR 2021 · similarity 0.82)
- MultiSports: A Multi-Person Video Dataset of Spatio-Temporally Localized Sports Actions (ICCV 2021 · similarity 0.82)
- ExAct: A Video-Language Benchmark for Expert Action Analysis (NeurIPS 2025 · similarity 0.81)
- ExpertAF: Expert Actionable Feedback from Video (CVPR 2025 · similarity 0.81)
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.
