MobiPhysio is a 2D video-based dataset for AI-based physiotherapy assessment, containing rehabilitation exercise videos from different camera angles with a wide range of movement variations. The videos are labeled and categorized by exercise type, with expert guidance from certified physiotherapists to ensure clinical relevance.
Use Cases
- Train computer vision models on labeled exercise videos to classify exercise type for automated assessment.
- Evaluate model robustness using videos from different camera angles to simulate real-life deployment scenarios.
- Develop AI systems capable of handling a wide range of movement variations present in the rehabilitation videos.
Strengths
- Dataset is designed with expert guidance from certified physiotherapists to ensure clinical relevance.
- Contains a wide range of movement variations to support training robust AI models.
- Videos are captured from different camera angles, providing multiple perspectives for analysis.
Limitations
- Dataset size, row count, and specific file formats are unknown, limiting reproducibility and scalability assessment.
- The absence of detailed column or feature descriptions makes precise data understanding and feature engineering difficult.
Provenance
- Source
- Harvard Dataverse
- Collection Method
- 2D video recordings of rehabilitation exercises, labeled and categorized with expert guidance.
- Time Range
- null
- Freshness
- null
- Geography
- null