Informace o publikaci
REHAB24-6: A multi-modal dataset of physical rehabilitation exercises
| Autoři | |
|---|---|
| Rok publikování | 2025 |
| Druh | Prototyp, poloprovoz, ověřená technologie |
| Fakulta / Pracoviště MU | |
| Citace | |
| Popis | REHAB24-6 is a multi-modal dataset of physical rehabilitation exercises designed to support the evaluation of human pose estimation methods and exercise feedback systems. The dataset contains 65 recordings of six common rehabilitation exercises performed by 10 subjects of varying age, sex, and fitness level, captured simultaneously by two RGB cameras and a 16-camera motion capture system. A physiotherapist supervised all recordings and instructed each subject to perform both correct and incorrect repetitions. The dataset provides RGB videos, 2D and 3D skeleton sequences, temporal segmentation of 1,072 individual repetitions, and binary correctness labels. The recordings cover diverse camera viewpoints, body types, and lighting conditions, offering a comprehensive testbed for exercise-correctness-related tasks. |
| Související projekty: |