Research map / Video / Action recognition
StrokeRehab: A Benchmark Dataset for Sub-second Action Identification
Aakash Kaku, Kangning Liu, Avinash Parnandi, Haresh Rengaraj Rajamohan, Kannan Venkataramanan, Anita Venkatesan, Audre Wirtanen, Natasha Pandit, Heidi Schambra, Carlos Fernandez-Granda
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 |
|---|---|---|---|
| replace MS-TCN segmentation with Seg2seq | MS-TCN · StrokeRehab stroke-patient test, video | action error rate: 0.392 -> 0.322 | Improved |
| replace MS-TCN with Raw2seq | MS-TCN · StrokeRehab stroke-patient test, sensor | action error rate: 0.330 -> 0.305 | Improved |
| use Raw2seq instead of Seg2seq on sensor data | Seg2seq · StrokeRehab stroke-patient test, sensor | action error rate: 0.337 -> 0.305 | Improved |
| add smoothing window to MS-TCN (ablation) | MS-TCN · StrokeRehab stroke-patient test, video | action error rate: 0.392 -> 0.390 | No clear effect |
| add boundary-detection refinement (ASRF over MS-TCN) (ablation) | MS-TCN · StrokeRehab stroke-patient test, video | action error rate: 0.392 -> 0.436 | Hurt |
| add smoothing window to MS-TCN (ablation) | MS-TCN · 50Salads | action error rate: 0.43 -> 0.32 | Improved |
| add smoothing window to MS-TCN (ablation) | MS-TCN · Breakfast | action error rate: 0.97 -> 0.51 | Improved |
| replace ASRF with Seg2seq | ASRF · Jigsaws | action error rate: 0.31 -> 0.17 | Improved |
| replace ASRF with Raw2seq | ASRF · Breakfast | action error rate: 0.45 -> 0.55 | Hurt |
| select model by validation AER instead of frame-wise accuracy (ablation) | MS-TCN · StrokeRehab stroke-patient test, video | action error rate: 0.408 -> 0.392 | Improved |
| train Raw2seq on stroke patients instead of healthy subjects (ablation) | Raw2seq · StrokeRehab stroke-patient test | action error rate: 0.405 -> 0.305 | Improved |
| train Raw2seq on stroke patients instead of healthy subjects (ablation) | Raw2seq · StrokeRehab severe-impairment test | action error rate: 0.819 -> 0.612 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MS-TCN — MS-TCN: Multi-Stage Temporal Convolutional Network for Action Segmentation (CVPR 2019) · better on 2 of 2 shared results
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
- PULSE: A Synchronized Five-Modality Dataset for Sensorimotor Coordination in Long-Horizon Daily Activities (NeurIPS 2026 · similarity 0.86)
- ActionAtlas: A VideoQA Benchmark for Domain-specialized Action Recognition (NeurIPS 2024 · similarity 0.83)
- CountLLM: Towards Generalizable Repetitive Action Counting via Large Language Model (CVPR 2025 · similarity 0.83)
- Recovering Complete Actions for Cross-dataset Skeleton Action Recognition (NeurIPS 2024 · similarity 0.83)
- VA-AR: Learning Velocity-Aware Action Representations with Mixture of Window Attention (AAAI 2025 · similarity 0.83)
- SeFAR: Semi-supervised Fine-grained Action Recognition with Temporal Perturbation and Learning Stabilization (AAAI 2025 · 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.
